Carbonate reservoir patterns play an important role in the production performance of oil and gas wells, and it is usually classified through static data analysis which cannot reflect the actual well performance. This paper takes the Tazhong No. 1 gas field in the Tarim Basin, China as an example to investigate the classification of carbonate reservoirs. The classification method mainly combines well test analysis with production analysis ― especially the Blasingame type curve method. Based on the characteristics of type curves for well test analysis and the Blasingame method, the relationship between the type curves and reservoir pattern was established. More than 20 wells were analyzed and the reservoirs were classifi ed into 3 major patterns with 7 sub-classes. Furthermore, the classifi cation results were validated by dynamic performance analysis of wells in the Tazhong No.1 gas field. On the basis of the classifi cation results, well stimulation (i.e. water flooding in a single well) was carried out in three volatile-oil wells, and the oil recovery increased by up to 20%.