Abstract: A new method of RCM reliability assessment for rail transit vehicle system maintenance decision\|making is proposed. The fault mechanism of rail transit vehicle system is complicated, and the influencing factors are redundant. A method based on the integration of RCM, PHM and data mining algorithms is proposed to construct a system maintenance decision\|making model. The difference with conventional method is that the proposed improved maintenance decision\|making method can define the maintenance model with higher accuracy and acquire the optimal maintenance interval of the system. The calculation is highly efficient and applicable for the reliability calculation of complex state system. This method can effectively reduce the under maintenance situation of conventional RCM, and reduce operation and maintenance costs, possessing certain promotion and application value.