Warm standby has been widely applied to power generation systems to ensure safe and reliable operation with lower energy consumption and shorter leading time compared with the case of hot standby and cold standby, respectively. Since there is a series of intermediate states between perfect functioning and complete failure for generating units, multi-state characteristics emerge in power generation systems. Considering different parameters of state transition time distributions before and after activation of standby units, as well as generating units with arbitrary state transition time distributions, this entails a need to propose a reliability evaluation technique to handle multi-state power generation systems considering warm standby with arbitrary state transition time distributions. An analytic method based on multi-state decision diagram (MSDD) is proposed in this paper for the reliability analysis of the proposed systems considering successful activation probabilities of warm standby units. The MSDD-based technique is applicable to multi-state power generation systems with arbitrary state transition time distributions for generating units. The proposed method enables the automated calculation of the system reliability and expected surplus performance. Furthermore, the optimum sequence of activating warm standby units that maximizes system reliability is formulated based on the proposed automatic reliability calculation method. Illustrative examples are provided to validate the effectiveness of the proposed technique.