Intensive research efforts have been dedicated to reliability analysis and optimization of standby systems. However, only few of the existing models allow a system element to be used/activated more than once during the mission time, and the existing work assumed a 1-out-of-n structure and fully reliable element activation procedures. This paper advances the state of the art by modeling and optimizing a m-out-of-n standby system subject to preventive replacements and imperfect element activations. Each element may be activated to execute the mission task and subsequently be preventively replaced multiple times during the mission; depending on the length of idle time, the element stays on standby or undergoes protection or maintenance actions prior to its next activation. Probabilistic approaches are suggested to evaluate mission success probability (MSP) and expected mission cost (EMC) for the considered standby system. Two optimization problems are solved, which determine the optimal operation and maintenance schedule (OMS) maximizing MSP and minimizing EMC, respectively. Effects of different model parameters on the MSP and optimization results are examined through examples. The proposed methodology is also extended to a generalized m-out-of-n standby system with heterogenous element performances, where the number of operating elements m may vary during the mission to meet a certain system demand.