The performance of an orthogonal frequency division multiplexing (OFDM) system is degraded if the peak-to-average power ratio (PAPR) is high. In general, in order to obtain optimal PAPR reduction using the partial transmitted sequence (PTS) technique, an exhaustive search of the possible subblocks and rotation factors must be done. As the number of subblocks and rotation factors increases, PAPR reduction improves, but the computational load becomes impractical. In order to reduce the complexity while still improving the OFDM system performance, a new method using a genetic algorithm (GA) is proposed to find a set of rotation factors that reduces both the PAPR and the computational load. A comparison is made between the proposed method and previously developed techniques such as exhaustive and gradient descent PTS methods. The superiority of the proposed method is demonstrated as a reduction in computational load compared with exhaustive PTS and the gradient method, and an improvement in performance compared with the iterative and gradient methods.