For exploring the feasibility that seed survival curve were transformed into line by survivor germination percentage transformed to survival rates, and to predict longevity of seed using seed survivor proportion ,Germination percent of wheat seeds of 8 seed lots of 3 species stored at 20℃,30℃and40℃ up to 810d with 3 levels of seed moisture content, were obtained. Seed survival curve which drawn using probability and survival proportion against storage time, fitting between regression equation of the probability and survival proportion established, and impact of interaction between storage temperature and seed moisture content on standard deviation of seed depth distribution in time were all compared and analyzed. The result showed that survivor proportion were easier than probability make seed survival curve to transform into line, the fitting of liner regression equation by survivor proportion against storage time established were also better than probability regression equation. Influence of interaction between storage temperature and seed moisture content on seed longevity were existing, its action were also bigger than independent temperature and moisture or logarithm of seed moisture content. Linear regress equation of predicting longevity of seed with survivor proportion, that is Gi/G0=A-P/10KE-CWT(t譵),were only proposed containing interaction between storage temperature and seed moisture content, and this equation were demonstrated to be fit to predict longevity of wheat seeds stored at low temperature genebank, by actual germination percent and storage time of common wheat seeds stored at low temperature genebank ,it was especially higher veracity for predicting germination percent than storage time.