By exchanging information between Turbo decoding and multi-user detector, an iterative algorithm for adaptive weighting parallel multi-user detection is proposed. The method is applied to space time block coded orthogonal frequency division multiplexing (STBC-OFDM) systems. Multi-user signals are separated using the algorithm, and the obtained soft bit information is used as the prior information for Turbo decoding. The extrinsic information obtained in Turbo decoding is then used as a priori information for weighting coefficients in the next iteration of adaptive weighting multi-user detection. Simulation results show that the proposed method converges rapidly in space time block coded OFDM systems. Compared with adaptive partial parallel multi-user detection, trail-and-error multi-user detection, and the conventional parallel interference cancellation multi-user detection, the proposed technique offers gains of 1.8 dB, 2.5 dB and 3.0 dB respectively.