To make the reliability estimation more practical and more accuracy, we proposed a method leverages two types of testing data to build the degradation model and reliability estimation. The corresponding artificial neural network training and inferencing the degradation process parameters are described. To enhance the accuracy of the degradation model, which is trained using both degradation testing data and life testing data, we describe the degradation process using a Gamma distribution. The parameters of Gamma process are set follow Gaussian distribution to describe the induvial difference and random effect. The parameters of Gaussian distribution given by moment estimation based on the training results. The accuracy of our proposed method is validated through a case study. The results indicate that our method offers distinct advantages in modeling the degradation process and in reliability estimation.