diff --git a/vllm_ascend/quantization/quant_config.py b/vllm_ascend/quantization/quant_config.py
index e358b253..99bd0499 100644
--- a/vllm_ascend/quantization/quant_config.py
+++ b/vllm_ascend/quantization/quant_config.py
@@ -221,6 +221,12 @@ packed_modules_model_mapping = {
         ["experts.0.gate_proj", "experts.0.up_proj", "experts.0.down_proj"],
         "fused_qkv_a_proj": ["q_a_proj", "kv_a_proj_with_mqa"]
     },
+    "pangu_ultra_moe": {
+        "gate_up_proj": ["gate_proj", "up_proj"],
+        "experts":
+        ["experts.0.gate_proj", "experts.0.up_proj", "experts.0.down_proj"],
+        "fused_qkv_a_proj": ["q_a_proj", "kv_a_proj_with_mqa"]
+    },
     "kimi_k2": {
         "gate_up_proj": ["gate_proj", "up_proj"],
         "experts":
@@ -241,6 +247,12 @@ packed_modules_model_mapping = {
         "experts":
         ["experts.0.gate_proj", "experts.0.up_proj", "experts.0.down_proj"]
     },
+    "pangu_ultra_moe_mtp": {
+        "gate_up_proj": ["gate_proj", "up_proj"],
+        "experts":
+        ["experts.0.gate_proj", "experts.0.up_proj", "experts.0.down_proj"],
+        "fused_qkv_a_proj": ["q_a_proj", "kv_a_proj_with_mqa"]
+    },
     "qwen3_next": {
         "qkv_proj": [
             "q_proj",
diff --git a/vllm_ascend/spec_decode/mtp_proposer.py b/vllm_ascend/spec_decode/mtp_proposer.py
index a152aa47..cd4b9a04 100644
--- a/vllm_ascend/spec_decode/mtp_proposer.py
+++ b/vllm_ascend/spec_decode/mtp_proposer.py
@@ -44,6 +44,8 @@ PADDING_SLOT_ID = -1
 _MTP_MODELS = {
     "DeepseekV3ForCausalLM":
     ("vllm.model_executor.models.deepseek_mtp", "DeepSeekMTP"),
+    "PanguUltraMoEForCausalLM":
+    ("vllm.model_executor.models.openpangu_mtp", "OpenPanguMTP"),
     "DeepseekV32ForCausalLM":
     ("vllm.model_executor.models.deepseek_mtp", "DeepSeekMTP"),
     "Qwen3NextForCausalLM":