diff --git a/python/cugraph-nx/cugraph_nx/interface.py b/python/cugraph-nx/cugraph_nx/interface.py
index fe492c43ca2..ccd8d418d30 100644
--- a/python/cugraph-nx/cugraph_nx/interface.py
+++ b/python/cugraph-nx/cugraph_nx/interface.py
@@ -22,6 +22,7 @@ class BackendInterface:
     @staticmethod
     def convert_from_nx(graph, *args, edge_attrs=None, weight=None, **kwargs):
         if weight is not None:
+            # MAINT: networkx 3.0, 3.1
             # For networkx 3.0 and 3.1 compatibility
             if edge_attrs is not None:
                 raise TypeError(
@@ -38,6 +39,13 @@ def convert_to_nx(obj, *, name: str | None = None):
 
     @staticmethod
     def on_start_tests(items):
+        """Modify pytest items after tests have been collected.
+
+        This is called during ``pytest_collection_modifyitems`` phase of pytest.
+        We use this to set `xfail` on tests we expect to fail. See:
+
+        https://docs.pytest.org/en/stable/reference/reference.html#std-hook-pytest_collection_modifyitems
+        """
         try:
             import pytest
         except ModuleNotFoundError:
@@ -51,17 +59,110 @@ def key(testpath):
                 return (testname, frozenset({classname, filename}))
             return (testname, frozenset({filename}))
 
-        string_attribute = "unable to handle string attributes"
+        no_weights = "weighted implementation not currently supported"
+        no_multigraph = "multigraphs not currently supported"
+
+        xfail = {}
+
+        from packaging.version import parse
 
-        skip = {
-            key("test_pajek.py:TestPajek.test_ignored_attribute"): string_attribute,
-            key(
-                "test_agraph.py:TestAGraph.test_no_warnings_raised"
-            ): "pytest.warn(None) deprecated",
-        }
+        nxver = parse(nx.__version__)
+        if nxver.major == 3 and nxver.minor in {0, 1}:
+            # MAINT: networkx 3.0, 3.1
+            xfail.update(
+                {
+                    key(
+                        "test_agraph.py:TestAGraph.test_no_warnings_raised"
+                    ): "pytest.warn(None) deprecated",
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedBetweennessCentrality.test_K5"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedBetweennessCentrality.test_P3_normalized"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedBetweennessCentrality.test_P3"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedBetweennessCentrality.test_krackhardt_kite_graph"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedBetweennessCentrality."
+                        "test_krackhardt_kite_graph_normalized"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedBetweennessCentrality."
+                        "test_florentine_families_graph"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedBetweennessCentrality.test_les_miserables_graph"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedBetweennessCentrality.test_ladder_graph"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedBetweennessCentrality.test_G"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedBetweennessCentrality.test_G2"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedBetweennessCentrality.test_G3"
+                    ): no_multigraph,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedBetweennessCentrality.test_G4"
+                    ): no_multigraph,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedEdgeBetweennessCentrality.test_K5"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedEdgeBetweennessCentrality.test_C4"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedEdgeBetweennessCentrality.test_P4"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedEdgeBetweennessCentrality.test_balanced_tree"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedEdgeBetweennessCentrality.test_weighted_graph"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedEdgeBetweennessCentrality."
+                        "test_normalized_weighted_graph"
+                    ): no_weights,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedEdgeBetweennessCentrality.test_weighted_multigraph"
+                    ): no_multigraph,
+                    key(
+                        "test_betweenness_centrality.py:"
+                        "TestWeightedEdgeBetweennessCentrality."
+                        "test_normalized_weighted_multigraph"
+                    ): no_multigraph,
+                }
+            )
         for item in items:
             kset = set(item.keywords)
-            for (test_name, keywords), reason in skip.items():
+            for (test_name, keywords), reason in xfail.items():
                 if item.name == test_name and keywords.issubset(kset):
                     item.add_marker(pytest.mark.xfail(reason=reason))
 
diff --git a/python/cugraph-nx/pyproject.toml b/python/cugraph-nx/pyproject.toml
index 8b0ae11fbe0..df1e7a7a9ab 100644
--- a/python/cugraph-nx/pyproject.toml
+++ b/python/cugraph-nx/pyproject.toml
@@ -40,6 +40,7 @@ test = [
     "pytest",
     "pytest-benchmark",
     "pytest-mpl",
+    "packaging >=21",
 ]
 
 [project.urls]
diff --git a/python/cugraph-nx/run_nx_tests.sh b/python/cugraph-nx/run_nx_tests.sh
index 5d3b616304d..8736d77010f 100755
--- a/python/cugraph-nx/run_nx_tests.sh
+++ b/python/cugraph-nx/run_nx_tests.sh
@@ -12,5 +12,9 @@
 #
 # Coverage of `cugraph_nx.algorithms` is reported and is a good sanity check that algorithms run.
 
-# NETWORKX_GRAPH_CONVERT=cugraph NETWORKX_BACKEND_TEST_EXHAUSTIVE=True pytest --pyargs networkx "$@"
-NETWORKX_TEST_BACKEND=cugraph NETWORKX_TEST_FALLBACK_TO_NX=True pytest --pyargs networkx --cov=cugraph_nx/algorithms --cov-report term-missing --no-cov-on-fail "$@"
+NETWORKX_GRAPH_CONVERT=cugraph NETWORKX_BACKEND_TEST_EXHAUSTIVE=True \
+NETWORKX_TEST_BACKEND=cugraph NETWORKX_TEST_FALLBACK_TO_NX=True \
+    pytest --pyargs networkx \
+    --cov=cugraph_nx/algorithms \
+    --cov-report term-missing --no-cov-on-fail \
+    "$@"