In order to investigate the general reliability assessment methods based on performance degradation data

two commonly used stochastic process approaches

bilinear process method and random-effect model were studied. Analyzing procedure and effectiveness of these two methodologies were studied and compared. Meanwhile

the two approaches were illustrated through practical applications. The residual plots and the 10th percentile curves of the two methods were presented to demonstrate the comparative results. The random-effect model yielded more volatile residuals and a lower and unsafe 10th percentile curve. Consequently the bilinear process model can be concluded to derive more reasonable results due to its one-stage estimation property.