In order to utilize the projection pursuit(PP)technique to scientifically and objectively evaluate the sustainable development capacity of regional water resources, an evaluation index system is established by selecting 20 indices based on the ecological environment, water resources and economic and social factors. By using the new bionic swarm intelligence algorithm, satin bower bird optimization(SBO)algorithm to optimize the best projection direction of PP technique, the evaluation model SBO-PP for the sustainable development capacity of water resources is proposed, and the gray wolf optimization(GWO)algorithm-PP model, the artificial bee colony(ABC)algorithm-PP model, the differential evolution(DE)algorithm-PP model, the biogeography-based optimization(BBO)algorithm-PP model, the particle swarm optimization(PSO)algorithm-PP model are compared, and the sustainable development capacities of water resources in Yunnan Province from 2006 to 2015 and in 2020 are evaluated. The results show that the SBO algorithm is better than the algorithms of GWO, ABC, DE, BBO and PSO in the optimization of PP technique, and it has better capability of solving precision, robust performance and extreme value. The sustainable development capacity of water resources in Yunnan from 2006 to 2007 evaluated by the SBO-PP model is worse, that from 2008 to 2015 is moderate, and that in 2020 is strong.