Marine information is an important way for us to know and study more about the ocean. Marine data makes the basic of marine information. Because of the huge quantity and diversity of marine data, and at the same time marine data is polyatomic variable, we start with statistical analysis methods to search for the regularity of the marine data. On one hand, we get the aggregate variation functions of the marine data by factor analyzing in aspect of the spatiality. Then we visually describe the marine status of the studied sea area with pre variogram function and post variogram function. On the other hand, we used cluster analysis method to get the verifying rule in time and make visible graphs of the marine data. In this way, we can also supply with the suggestions in classifying the sea seawater quality. The data processing result shows that the suggested methods in this article are both operable and effective. At the same time some reasonable suggestions are given in the article.