Generalized frequency division multiplexing (GFDM) is proposed as a candidate waveform to tackle new challenges posed on the physical layer for the fifth generation of wireless communication systems. In this paper, we propose a highly efficient algorithm for GFDM data detection on the basis of orthogonal approximate message passing (OAMP). We further combine with the conjugate gradient method and the approximate diagonalization of the equivalent channel matrix to significantly reduce the complexity engendered by matrix inverse operations in OAMP. Numerical simulations show that the proposed algorithm performs closely similar to the optimal performance at reduced computational costs in GFDM signal detection.