diff --git a/src/google/adk/evaluation/simulation/per_turn_user_simulator_quality_v1.py b/src/google/adk/evaluation/simulation/per_turn_user_simulator_quality_v1.py
index 239fa31a71..8d2a88dda4 100644
--- a/src/google/adk/evaluation/simulation/per_turn_user_simulator_quality_v1.py
+++ b/src/google/adk/evaluation/simulation/per_turn_user_simulator_quality_v1.py
@@ -302,9 +302,15 @@ def _evaluate_first_turn(
           eval_status=EvalStatus.NOT_EVALUATED,
       )
 
+    user_text = get_text_from_content(first_invocation.user_content)
+    if user_text is None:
+      return PerInvocationResult(
+          actual_invocation=first_invocation,
+          eval_status=EvalStatus.NOT_EVALUATED,
+      )
+
     score = int(
-        get_text_from_content(first_invocation.user_content).strip()
-        == conversation_scenario.starting_prompt.strip()
+        user_text.strip() == conversation_scenario.starting_prompt.strip()
     )
     return PerInvocationResult(
         actual_invocation=first_invocation,
diff --git a/tests/unittests/evaluation/simulation/test_per_turn_user_simulation_quality_v1.py b/tests/unittests/evaluation/simulation/test_per_turn_user_simulation_quality_v1.py
index be0a0394c5..296b9935c6 100644
--- a/tests/unittests/evaluation/simulation/test_per_turn_user_simulation_quality_v1.py
+++ b/tests/unittests/evaluation/simulation/test_per_turn_user_simulation_quality_v1.py
@@ -514,6 +514,32 @@ def test_evaluate_first_turn_failure():
   assert result.eval_status == EvalStatus.FAILED
 
 
+@pytest.mark.parametrize(
+    "user_content",
+    [
+        types.Content(role="user", parts=[]),
+        types.Content(),
+    ],
+)
+def test_evaluate_first_turn_not_evaluated_when_user_content_has_no_text(
+    user_content,
+):
+  """First turn with empty/None parts is not evaluated instead of crashing."""
+  evaluator = _create_test_evaluator(
+      threshold=1.0, stop_signal="test stop signal"
+  )
+  conversation_scenario = _create_test_conversation_scenario(
+      conversation_plan="plan",
+      starting_prompt="test starting prompt",
+  )
+  invocation = Invocation(invocation_id="1", user_content=user_content)
+
+  result = evaluator._evaluate_first_turn(invocation, conversation_scenario)
+
+  assert result.score is None
+  assert result.eval_status == EvalStatus.NOT_EVALUATED
+
+
 def test_aggregate_conversation_results_all_pass_produces_pass():
   evaluator = _create_test_evaluator()
   results = [