diff --git a/pkg/scheduler/backend/queue/active_queue.go b/pkg/scheduler/backend/queue/active_queue.go
index fa1bca28368..0fdd0f0d450 100644
--- a/pkg/scheduler/backend/queue/active_queue.go
+++ b/pkg/scheduler/backend/queue/active_queue.go
@@ -323,14 +323,13 @@ func (aq *activeQueue) unlockedPop(logger klog.Logger) (*framework.QueuedPodInfo
 	}
 	aq.schedCycle++
 
-	// Update metrics and reset the set of pending plugins for the next attempt.
-	// Note: We don't clear UnschedulablePlugins here because:
-	// 1. If the pod schedules successfully, we need UnschedulablePlugins for logging/debugging
-	// 2. If the pod fails to schedule again, UnschedulablePlugins will be overwritten anyway
+	// Update metrics for unschedulable plugins.
+	// Note: We don't clear UnschedulablePlugins and PendingPlugins here because:
+	// 1. If the pod schedules successfully, we need them for logging/debugging
+	// 2. If the pod fails to schedule, they will be cleared and repopulated in handleSchedulingFailure
 	for plugin := range pInfo.UnschedulablePlugins.Union(pInfo.PendingPlugins) {
 		metrics.UnschedulableReason(plugin, pInfo.Pod.Spec.SchedulerName).Dec()
 	}
-	pInfo.PendingPlugins.Clear()
 	pInfo.GatingPlugin = ""
 	pInfo.GatingPluginEvents = nil
 
diff --git a/pkg/scheduler/backend/queue/scheduling_queue_test.go b/pkg/scheduler/backend/queue/scheduling_queue_test.go
index 585ab56ef8a..5fec77b1f23 100644
--- a/pkg/scheduler/backend/queue/scheduling_queue_test.go
+++ b/pkg/scheduler/backend/queue/scheduling_queue_test.go
@@ -649,7 +649,7 @@ func Test_InFlightPods(t *testing.T) {
 			},
 		},
 		{
-			name:                         "popped pod preserves UnschedulablePlugins but clears PendingPlugins",
+			name:                         "popped pod preserves UnschedulablePlugins and PendingPlugins",
 			isSchedulingQueueHintEnabled: true,
 			initialPods:                  []*v1.Pod{pod1},
 			actions: []action{
@@ -670,9 +670,9 @@ func Test_InFlightPods(t *testing.T) {
 					if !poppedPod.UnschedulablePlugins.Equal(sets.New("fooPlugin2")) {
 						t.Errorf("QueuedPodInfo from Pop should preserve UnschedulablePlugins, expected fooPlugin2, got: %+v", poppedPod.UnschedulablePlugins)
 					}
-					// PendingPlugins should still be cleared
-					if len(poppedPod.PendingPlugins) > 0 {
-						t.Errorf("QueuedPodInfo from Pop should have empty PendingPlugins, got instead: %+v", poppedPod.PendingPlugins)
+					// PendingPlugins are preserved after Pop() for logging
+					if !poppedPod.PendingPlugins.Equal(sets.New("fooPlugin1")) {
+						t.Errorf("QueuedPodInfo from Pop should preserve PendingPlugins, expected fooPlugin1, got: %+v", poppedPod.PendingPlugins)
 					}
 				}},
 				{callback: func(t *testing.T, q *PriorityQueue) {
@@ -940,8 +940,10 @@ func TestPop(t *testing.T) {
 
 			// Now check result of Pop.
 			poppedPod = popPod(t, logger, q, pod)
-			if len(poppedPod.PendingPlugins) > 0 {
-				t.Errorf("QueuedPodInfo from Pop should have empty PendingPlugins, got instead: %+v", poppedPod)
+			// PendingPlugins are preserved after Pop() so they can be logged if scheduling
+			// succeeds, or cleared in handleSchedulingFailure() if it fails.
+			if len(poppedPod.PendingPlugins) != 1 || !poppedPod.PendingPlugins.Has("fooPlugin1") {
+				t.Errorf("QueuedPodInfo from Pop should preserve PendingPlugins, expected fooPlugin1, got instead: %+v", poppedPod)
 			}
 		})
 	}
diff --git a/pkg/scheduler/schedule_one.go b/pkg/scheduler/schedule_one.go
index 4908b711da7..1160f456b8f 100644
--- a/pkg/scheduler/schedule_one.go
+++ b/pkg/scheduler/schedule_one.go
@@ -342,7 +342,7 @@ func (sched *Scheduler) bindingCycle(
 	}
 	// Count pods scheduled after being flushed from unschedulablePods
 	if assumedPodInfo.WasFlushedFromUnschedulable {
-		logger.Info("Pod scheduled after flush from unschedulablePods", "pod", klog.KObj(assumedPodInfo.Pod), "unschedulablePlugins", assumedPodInfo.UnschedulablePlugins)
+		logger.Info("Pod scheduled after flush from unschedulablePods", "pod", klog.KObj(assumedPodInfo.Pod), "unschedulablePlugins", assumedPodInfo.UnschedulablePlugins, "pendingPlugins", assumedPodInfo.PendingPlugins)
 		metrics.PodScheduledAfterFlush.Inc()
 	}
 	// Run "postbind" plugins.
@@ -1063,6 +1063,11 @@ func (sched *Scheduler) handleSchedulingFailure(ctx context.Context, fwk framewo
 	err := status.AsError()
 	errMsg := status.Message()
 
+	// Clear plugin sets to avoid stale data from previous scheduling attempts.
+	// They will be repopulated below for FitError cases.
+	podInfo.UnschedulablePlugins.Clear()
+	podInfo.PendingPlugins.Clear()
+
 	if err == ErrNoNodesAvailable {
 		logger.V(2).Info("Unable to schedule pod; no nodes are registered to the cluster; waiting", "pod", klog.KObj(pod))
 	} else if fitError, ok := err.(*framework.FitError); ok { // Inject UnschedulablePlugins to PodInfo, which will be used later for moving Pods between queues efficiently.