Gas content is the key parameter of shale gas resource evaluation and favorable area optimization. The main factors controlling gas content need to be considered in exploration. But due to the fact that not each sample can cover all factors, samples which have all kinds of experimental data are rare and the data utilization rate is poor when using common methods in finding out the main controlling factors. Combining with shale gas occurrence mechanisms and all kinds of experimental data, twelve factors that may control the total gas content are selected from geochemical parameters and reservoir properties. By using grey model correlation analysis for the chaos system where sample data does not cover all factors, Deng grey correlation and absolute grey correlation are calculated. we multiplied both together to form a similar grey correlation, and then establish the total gas content prediction model accordingly. Compared with principal component analysis and euclidean distance analysis, we used the linear backward regression method to validate the pros and cons of possible results. This shows that the main controlling factors shown by grey correlation analysis are clay content, S1 + S2, porosity and TOC. The R2 of the linear model is 0.878, which is higher than the other two methods. A gas content distribution map of the study area shows that the trend agrees with the actual well tests.