In order to study the effectiveness and feasibility of the gravitational search algorithm (GSA) - Projection Pursuit (PP) model applying to evaluate the carrying capacity of water resources, the regional water resources carrying capacity evaluation index system and classification standards were constructed by selecting 14 indicators from the water system, economic and social system, and water environmental system, the best projection direction was optimized using GSA algorithm, and the GSA-PP evaluation model for water resources carrying capacity was also proposed, then the PP models established by particle swarm optimization (PSO) algorithm, differential evolution (DE) algorithm and cuckoo search (CS) algorithm optimization were respectively compared with the GSA-PP model, finally the water resources carrying capacity in 8 Wenshan counties was evaluated as case studies. The results showed that the fitness value of 10 consecutive optimizations by GSA algorithm is 1147.9363, they were all better than algorithms of PSO, DE and HS, it has higher solution accuracy and stable performance. The evaluation results of GSA-PP model in Malipo County, Maguan County and Wenshan city were“may carrying capacity”, the evaluation results of the remaining five counties as “basically may carrying capacity”. The evaluated and sorted results of GSA-PP were similar to the DE-PP model; and the evaluation results of GSA-PP were also same as PSO-PP and CS-PP model, but there were differences in the sorted results. Therefore, the solution accuracy of the intelligent algorithm determines the pros and cons of evaluation effects.