Assessment of water quality by firefly algorithm based on BP neural network model ( FA-BP model) is built. In this model, the ev aluation index function is constructed by BP Artificial Neural Network Algorithm ( BP model) , and Firefly Algorithm ( FA model) is intr oduced to optimize weight values and thresholds to find the optimal solution. Fuzzy Comprehensive Evaluation method, Grey Incidence Analysis Algorithm and FA-BP model will be applied to evaluate the water quality of the five main rivers in Liany ungang City including Longwei, Yudai, Dapu, Paidan, and Dong yan River. The results show that the Fuzzy Comprehensive Evaluation method is difficult to use for slight pollution rivers with several slightly over standard indexes. It will be easy to ignore the impact of ex treme indexes by Grey Incidence Analysis Algorithm. FA-BP model solves the shortcomings of the two methods. The evaluation results pr ovide an important reference for the formulation of reasonable measures. It is a relatively comprehensive evaluation method and has agood a pplication prospect in water qualit yevaluation.