In reliability-based design optimization, the constraints consider the probability of the satisfaction/failure of critical events. Lately, reliability-based design optimization has been applied to topology optimization, resulting in the development of reliability-based topology optimization. And though reliability-based topology optimization can be a useful and meaningful method, it requires excessive computational resources. Therefore, this research proposes a parallel-computed reliability-based topology optimization using the response surface method. This paper demonstrates that the proposed method greatly reduces the computation requirement of reliability-based topology optimization. The proposed methodology is then applied to design microelectromechanical systems. Specifically, in microelectromechanical systems, reliability-based topology optimization can be highly effective because of randomness generated during the etching process and scaling effect. The proposed method successfully designs new devices and verifies the designs via experiment.