This paper focuses on the automotive component failure rate detection using Weibull analysis and other survival analysis techniques. Detailed attention is paid to three areas: 1) overall failure rates are described statistically first, and data cleaning and definitions of ‘failed’ and ‘censored’ data within the research time or warranty period are made; 2) Kaplan-Meier life curves and Log-rank tests are used to compare the component reliability over time and explore risks factor related to the component failure; 3) Weibull regressions, with two and three parameters, are applied to fit real-world reliability data from different test conditions, and to predict the automotive component failure trend over future time. The analysis results agree well with real-world test data, and provide reasonable prediction of future failure trends.