The linear Minimum Mean Square Error (MMSE) signal detection algorithm in the massive Multiple-Input Multiple-Output (MIMO) system is difficult to use in practical engineering due to the complex inversion of the high-dimensional matrix of the signal detection algorithm. In order to solve this problem, the article proposes a low-complexity hybrid iterative algorithm based on the idea of matrix block and combined with Neumann series expansion algorithm. Using the Neumann series second-order expansion of the inverse matrix of the weighted matrix in the MMSE algorithm as the initial value of the iterative inversion of the block matrix can effectively improve the convergence speed of the algorithm. The simulation results show that this method can have similar performance as the traditional MMSE algorithm with fewer iterations and lower complexity.