These days, automotive industries are facing intensive competition. New products should be delivered into market as soon as possible. On the other hand, customers also require new products having good reliability. Laboratory life testing is the traditional way to examine the reliability of products. Sample products should be tested till failures, which usually involve tremendous cost and time. Obviously, the reliability testing process extends the duration of 'design-to-market'. Cutting down the testing process can reduce the producing cycle properly. Currently, the accelerated life test (ALT) has been widely used in many companies as an alternative method. Although ALT reduces the cost of reliability testing through applying more severe environmental conditions than the normal ones on products, it is no longer sufficient as it does not describe the process of products' failure explicitly and it is still highly dependent on physical testing. Consequently, prediction models need to be developed for better understanding of the products' reliability.