The prediction of oilgas reservoir thickness is a key technology for oilgas reservoir description. Nevertheless, the relationship between the thickness of the reservoir and the seismic information is nonlinear, and hence the present mathematical method can hardly perform the fitting. To solve this problem, the author used a new mathematical statistical method, namely  Support Vector Machine Regression. This paper described the Support Vector Machine Regression and the corresponding analytical methods based on the principle of this technology. On such a basis and with the practical data as an example, the authors successfully predicted the reservoir parameter—the thickness. It is concluded that the technology is feasible and practical.