Abstract: Squid (Sthenoteuthis oualaniensis), one of the important economic Cephalopods in the South China Sea, has great development potential and economic value, playing an increasingly significant role in the marine fisheries of the South China Sea. In order to understand the variation of catch per unit effort (CPUE) and make scientific predictions, we constructed a gray prediction model GM (1,N) and combined it with marine environmental factors, then fitted and predicted the production data of the South China Sea squid from spring to summer of 2013−2019. Besides, we modified the background value of the original model by Simpson formula and Fourier series, and corrected the residual error of the simulated value. The results show that the average relative errors of the traditional GM (1,N) model and the optimized grey GM (1,N) model were 7.78% and 2.54%, respectively. For the prediction of squid CPUE in 2019, the optimized grey GM (1,N) model reduced the relative error from 4.79% of the traditional GM (1,N) model to 1.87%. It is showed that the optimized grey system model has a higher prediction accuracy than the traditional GM (1,N) model, which provides a new idea on how to predict the relative abundance of squid resources accurately.