User scheduling in mmWave multi-user multiple-input multiple output (MU-MIMO) systems involves the joint optimization of user selection and beam pair selection due to its hybrid beamforming architecture. Most of the existing solutions are based on ideal channel state information (CSI). How to choose the users to be scheduled and their beams for operation in the case of non-ideal CSI is still challenging. In this study, we propose a blind beam search assisted user scheduling (BSUS) algorithm operating in the absence of ideal CSI. The proposed method consists of two steps. In the first step, the beam information acquired in the initial beam training is used to identify the channels’ orthogonality and classify the users into groups. In the second step, a scheduled user is chosen from each group to maximize the equivalent channel energy. Subsequently, the selection of beam pairs for communication is simplified owing to the information acquired in initial beam training. Simulation results show that the proposed scheme can achieve 93.6% of the maximal sum rate, and maintain a low computational complexity.