Based on the high dimension of the channel matrix and the complexity of the received signals, a hybrid iterative algorithm signal detection on uplink for large-scale MIMO systems is proposed, which combines adaptive damped Jacobi (DJ) algorithm and conjugate gradient (CG) algorithm. Firstly, conjugate gradient algorithm is used to provide effective search direction for adaptive damped Jacobian iterative algorithm. Then, Chebyshev method is proposed to eliminate the influence of relaxation parameters on signal detection to reduce the complexity of the algorithm and accelerate the convergence speed. Finally, the soft information is approximately solved by using the bit likelihood ratio in channel coding and decoding. The simulation results show that the hybrid iterative algorithm converges quickly and approximately achieves the best MMSE detection performance under a small number of iterations, and the algorithm complexity is far lower than that of MMSE algorithm.