The multiple-input multiple-output-orthogonal frequency division multiplexing (MIMO-OFDM) system has huge advantages of high data rates and good performance in wireless applications; however, it has a major disadvantage, which is the high peak-to-average power ratio (PAPR), that distorts the signal quality and reduces the overall efficiency of the system. A new method to minimize the PAPR in MIMO-OFDM systems has been developed in this paper, by incorporating an enhanced fireworks algorithm (EFWA) with partial transmit sequence (PTS) method. This is the case of traditional PTS methods that aid in reducing the PAPR but normally have a high computational complexity and therefore are not ideal for the real-world application. The proposed method employs the EFWA, an advanced variant of the fireworks algorithm, to minimize the phase factors in an efficient manner apart from any computational burden. Thus, the EFWA has incorporated the following realizations: the minimal explosion amplitude operator, the new explosion sparks operator, and the new mapping strategy which can also increase the speed of convergence and robustness of the optimization. The results obtained for the proposed PTS scheme with EFWA offer a better PAPR with 2.98 dB than the conventional methods, namely, prime learning ant lion optimization (PL-ALO) and moth flame optimization with improved firefly algorithm (MFO-IFFA).