As one of core components of machine tool,the reliability of spindle is very important for improving machine tool quality. According to spindle performance degradation mechanism analysis,the acoustic emission signal is used to monitor and study the performance state of spindle. A reliability evaluation method is proposed for the long-life and less failure sample problem based on machine tool degradation path models. At last,this method is used to solve the machine tool spindle reliability assessment,which integrates virtual augmented sample method. It effectiveness and practicability is proved through experiment results.