Abstract: In the karst areas of eight provinces in southwestern of China, the number of underground river and their total discharge are two important parameters that draw mauch attention. The statistics in the 1980s yielded 2 836 underground rivers and total discharge 1 482 m3/s. Whereas in the 1990s, the statistics for all the 8 provinces, corrected by the database in recent years, shows that there are 2 523 underground rivers with total discharge about 1 321.7 m3/s. Of them, there are 120 underground rivers with single discharge more than 2 000 L/s, and their discharge sums 699.7 m3/s accounting for 52.94% of the total. The underground rivers with discharge between 50-500 L/s is the most, occupying 51.55% of the total; next is those less than 50 L/s which is 723 in number and accounts for 22.43% of the total. This paper points out that the statistics above are problematic with over-estimation in karst underground river discharge and large-discharge karst springs. Taking two examples of the Daxiaojing and Zhaidi underground river systems, this paper presents the specific amount of double counting. Based on the old 1∶200 000 hydrogeological survey data, the total re-statistics account for 57.1% and 7.1% separately for the total outlet discharge of the Daxiaojing and Zhaidi river systems, respectively. While these values can be increased separately to 134.3% and 86.1% using the latest survey data. It can concluded further that with increasing coverage of 1∶50 000 hydrogeological surveys, more outlets will be counted, the amount of the bouble counting by the traditional methods will be greater. The primary reason for such an overestimation problem lies in mess of the underground river system and sub-system. Finally, this paper discusses the spatial attributes of karst underground river systems and proposes countermeasures to metigatye the problem of double counting. It would help establish realistic statistics of discharge of underground rivers and major springs of karst areas in southwestern China. 