A segmented nonlinear accelerated degradation model (SNADM) was introduced based on the equivalent method of accumulative damage theory, which tackles the problem that product life is difficult to determine with degradation rate being a function of the variable of time. A segmented expression of the function of overall accumulative degradation was derived and combined with nonlinear function and accelerated degradation function, from which a expression of SNADM was obtained. The parameters of the SNADM were identified by numerical iteration; further the statistical function of degradation track was extrapolated. The reliability function was determined through the type of random process of the degradation distribution. Then an evaluation to product storage life was undertaken by combining the statistical function of degradation track, reliability function and failure threshold value. An example of a missile refrigerator undergoes a step-down stress accelerated degradation test while test data were evaluated with SNADM. The technology introduced is validated with the resultant coincidence of both evaluated and field storage lives.