In this letter, we focus on the high peak to average power ratio (PAPR) reduction problem in tone reservation-based orthogonal frequency division multiplexing systems. We first propose a parallel Tabu search (PTS)-based scheme to find a sub-optimal peak reduction tone (PRT) set. After finding the sub-optimal PRT set, we apply it in an adaptive iterative clipping and filtering (AICF) method for PAPR reduction. Furthermore, PAPR reduction and bit error rate (BER) performances are compared among the AICF method, the adaptive scaling, the adaptive amplitude clipping and the fast iterative shrinkage-thresholding algorithm schemes. Simulation results verify that the PTS-based PRT scheme can obtain better secondary peaks with lower computational complexity, and the AICF scheme effectively reduces PAPR with a faster convergence speed while its BER performance is only slightly worse than for the existing methods.