Non-orthogonal multiple-access (NOMA) scheme using massive multiple-input multiple-output (MIMO) transmission suffers from interference between users. In this paper, we propose a novel MIMO-NOMA framework based on space-time block code (STBC) to decrease the multiple-access interference (MAI), when the number of users is increasing. In this paper, besides giving an analysis of network, cluster, group, and users, we first present the system model. Then, an efficient precoder and detector are designed based on interference analysis. In particular, the proposed scheme can decompose a massive MIMO-NOMA system into multiple separated single-input single-output NOMA channels. Also, analytical results are developed to evaluate the performance of the system. To address the dramatic increment of power level, with the aid of the proposed algorithm, the optimal solution is achieved. Mathematical analyses for both MIMO-NOMA and STBC are presented to design efficient precoding and detection. Finally, simulation results, based on STBC, are provided to verify the theoretical analysis to demonstrate significant capacity improvements of the massive MIMO-NOMA.