Orthogonal frequency division multiplexing (OFDM) has a very high peak-to-average power ratio (PAPR) that causes a severe nonlinear distortion in practical hardware implementation of high power amplifiers (HPA). In this article, a new PAPR reduction method is proposed based on autoregressive (AR) error filtering. This method proposes the use of signal whitening property of error filtering as a preprocessing step to remove the predictable content of stationary stochastic processes which can reduce the autocorrelation of input data sequences and is shown to be a very effective solution for the PAPR problem in OFDM systems. It is shown that the proposed method can achieve a significant reduction in PAPR without degrading the error performance or power spectral levels. It is also shown that the proposed method is applicable to any modulation scheme and can work for any number of subcarriers under both additive white Gaussian noise and wireless Rayleigh fading channel.