Ommastrephes bartramii is a kind of short-lived species which is one of the economic cephalopods in the Northwest Pacific. Optimizing the resource abundance prediction model can provide a scientific and effective basis for fishery production. This study used the fishing data of neon flying squid from 1998 to 2016. Firstly,GM (1,1) models are established for resource abundance (CPUE) sequences of different time lengths. The CPUE sequence with the smallest relative error and variance is selected to perform grey correlation analysis with the environment and climate factors of the spawning and fattening grounds, including Pacific Decadal Oscillation Index (PDO), average sea surface temperature at spawning ground (SGSST), average sea surface temperature at fattening grounds (FGSST), average chlorophyll concentration at spawning ground(SGC), average chlorophyll concentration at fattening ground (FGC)to evaluate the importance of environmental factors. And based on the results, we established 6 grey prediction models of different orders[GM (0, N) model and GM (1, N) model]. Finally,we selected the model with relatively small fitting errors and prediction errors as the best model to predict the abundance of neon flying squid resources. The results show that the average relative error of the GM (1,1) model of the 8-year CPUE sequence is the smallest (6.28%). The prediction accuracy of the GM (0, N) models is generally higher than that of the GM (1, N) models. The GM (0,5) model 4 which included SGSST in February, FGSST in October, FGC in August, and PDO in October have the best model effects. Its relative error of fitting is 3.87%, and the relative error of prediction is 1.18%.Therefore,we suggested that this model can be used to forecast the resource abundance of neon flying squid.