To predict seed longevity and viability of Sorghun vulgare, Agopyrum escutentum, Lactuca sativa, and Amaranthus by survivor proportion and main storage condition factors, seeds of 8 seed lot of Sorghun vulgare and Amaranthus with 3 levels of water content were stored at 20℃, 30℃ and 40℃ for lasting 810 d. Data sets of germination percent and storage period of seed from storage experiments of Sorghun vulgare and Amaranthus seed lot and from literature reported for Fagopyrum escutentum and Lactuca sativa seed lot were used to estimate mean viability period (p50) and standard deviation of seeds death distribution in time (σ) for different seed lot using regression equations of probability and survivor proportion in linearity, and fitting of the equations were also synchronously compared. Moreover, p50 and σ by two equations estimated were used to analyze storage condition factors with stepwise regression method, including storage temperature(t℃)and its quadratic term（t2), water content of seed (m%)and its logarithm(logm), and interaction between storage temperature and water content of seeds (t×m and t×logm), respectively. The results showed that estimated p50 and σ with survivor proportion to estimate was feasible and had more statistical significance than probability estimated. Seed viability was quickly lost and its longevity was decreased by interaction between storage temperature and water content of seed (t×m). The mathematic model Gi/G0=A-P/10K-CWT (t×m) containing the interaction factor and survivor proportion had higher accuracy to predict initial germination percent of Lactuca sativa than the model Vi=Ki- p/10 K-CWT (t×m). Thus, the former and latter models were recommended to predict initial germination rate with longevity and viability of seed known and unknown, respectively.