With the development of the network, online education breaks the defects of time and space, and becomes a way that more and more students choose to learn. This paper combines grey correlation analysis, optimizes BP neural network using PSO particle algorithm, and constructs GRA-PSO-BP model. The initial education evaluation indexes are improved by this intelligent algorithm model to construct an online education quality evaluation system, and then the optimized index system is used as a guide to evaluating the online education quality using this model. The results show that: the dispersion of the data of each index is