Reliability Demonstration Testing (RDT) for non-repairable systems has been successfully implemented in many industries, such as microelectronics, aerospace and healthcare. In designing RDTs, the famous beta-binomial formula is often applied to determine the necessary sample size so that the required reliability metric at a given confidence level can be demonstrated. However, many systems, such as cars and combine harvesters, are repairable systems. For repairable systems, RDT design methods have been developed mainly for Homogeneous Poisson Processes (HPP). This paper provides a new RDT design method applicable for both HPP and Non-Homogeneous Poisson Process (NHPP) cases. The result shows that the proposed method can effectively determine the required testing time and sample size of test units to demonstrate the required Mean Time Between Failures (MTBF). A comparison study is also conducted to illustrate the superior performance of the proposed method over some of the existing methods.