Massive MIMO is an extension of MIMO (Multiple-input Multiple-output), which essentially groups together several number of transmitting antennas and receiving antennas in order to provide better spectrum efficiency and better throughput. Massive MIMO technology will almost be a key component of the super-fast 5G networks of the future generation. However, the Multiple User Interference (MUL) is the major problem that needs to be address to detect each user appropriately at the base station. Hence, detection of encoded signals is one of the crucial receiver functions in MIMO wireless communication. The well-known detection algorithms like Zero forcing detector (ZF) and MMSE provides less complex design at the cost of reduced performance. Maximum Likelihood detector (ML) gives the optimal performance, but its computational complexity limits its usage in practical systems. Thus, to balance the trade-off between complexity and performance, the suboptimal Likelihood Ascent Search (LAS) algorithm has been proposed. However, the LAS algorithm detection performance depends on selection of initial solution. So, we propose Neighborhood search algorithm to select the initial vector and this healthy initial vector can be used for LAS algorithm to find solution. This algorithm is iterative in nature and it searches for that vector which minimizes the maximum likelihood (ML) cost in the neighborhood. The proposed algorithm is expected to balance the tradeoff between performance and complexity of design. © 2019 IEEE.