With the development of science and technology and improvement of material, the reliability of industry products becomes increasingly higher. Some life tests result in few or no failures. In such cases, it is difficult to assess reliability with traditional life tests which only record time-to-failure. For some products, it is possible to obtain degradation measurements over time and these measurements may contain useful information about product reliability. In this paper, statistical models and its inference methods for degradation measurements have been studied systematically. The contributions of this paper are as follows.The concepts of degradation failure and the framework of reliability analysis from degradation measurements have been discussed. The inference methods for the repeated-measures degradation data are proposed.The performance degradation is caused by interior damage cumulative processes of products that may be discrete or continuous. When the performance degradation is caused by the discrete cumulative damages, the degradation failure model based on renewal process is proposed and used to analyze the reliability of metallized film pulse capacitors in the laser device. When the degradation process is continuous, two models based on Wiener-Einstein and Gamma processes are presented respectively. Because it is difficult to obtain an analytical expression of the failure model using the latter method, we present a failure model based on Monte-Carlo simulation. When the degradation process is composite one of discrete and continuous damage cumulative processes, a degradation failure analytical model is presented based on diffusion process.Competing risk involving multiple failures are becoming increasingly common and important in practice. A general model involving both traumatic and soft degradation failures is proposed in which the dependence of traumatic failure intensities on the degradation level is included. The inference methods of unknown parameters of the competing failure model are presented based on proportional hazard model and location-scale model.The degradation failure standard may be random to some products. There are two kinds of degradation failure with random failure standard, the first kind is the degradation failure with relative failure standard that is determined by the random initialized performance measurements. A degradation failure model is proposed considering the relative failure standard. The second kind of degradation failure of random failure standard is the stress-strength interference failure considering stochastic stress and strength aging degradation. Two SSI reliability analysis methods considering strength aging degradation with cycle stochastic stress and stochastic