The paper introduced the idea of chaos optimization into differential optimization algorithms so as to form chaotic differential algorithm, and applied it to the function optimization problems of model parameter of river water quality. The results show that the chaotic differential algorithm is better in solving the parameter problem than the differential algorithm on accuracy and time. It brought the chaos optimization ideas of ergodicity and randomness into differential optimization algorithm, and carried out chaotic detail search near the optimal location after running difference method in each iteration, cooperated with special termination criterion to look for optimization. The method significantly shortened the calculation time of chaotic search , overcame the precocious defects of difference algorithm in the later period and improved the convergence speed and accuracy of the model.