In this paper, massive MIMO and NOMA techniques are jointly considered in a two tier heterogeneous network for accommodating huge traffic. In addition, full duplex communication is used at SBS to enhance system performance. First, SINR and rate equations are modeled while considering imperfect channel state information and zero-forcing beamforming. In the proposed framework, both macro base station and small cell base station can serve multiple users by using NOMA. Since, existing literature mainly focuses on orthogonal multiple access (OMA) based UA schemes, it is necessary to re-design user association (UA) schemes for NOMA. Hence, based on the proposed framework, a joint UA and power allocation (PA) optimization problem is designed with the perspective of optimizing the sum-rate of the system while accounting for quality of service and backhaul constraints. Since, the initially formulated problem is non-convex, a distributive algorithm is propounded where a joint problem is separated into two convex subproblems (viz., UA and PA). Furthermore, the complexity of the proposed algorithm is examined and Jain’s fairness index is measured. Simulation results demonstrate the effectiveness of the proposed algorithm by comparing with optimal solution and other base-line algorithms.