Abstract: The numerical forecasting of monthly-scale regional climate change is tested with the CSU-RAMS mesoscale numerical model nested in 9-level general circulation spectral model(GCM).The results show that the GCM assembly forecast has reproduced realistic large-scale average circulation;based on this,the CSU-RAMS model nested in GCM reproduces regional circulation and its short-range climate change.The "regional climate model" of nesting a meso-scale model into GCM is a valid method for forecasting regional short-range climate change.