Massive multiple input multiple output(MIMO)-based fully-digital receive antenna arrays eventuate a huge amount of circuit costs and complexity to direction of arrival(DOA) estimation, which is hard to satisfy the needs of high precision and low cost in future green wireless communication. To address this challenge, a novel heterogeneous hybrid MIMO receiver is proposed in this paper and a high performance DOA estimator called heterogeneous cross-minimum distance (HCMD) is developed based on the structure. The antenna arrays are first divided into multiple groups, and each group adopts a different hybrid structure. The virtual antenna arrays of these groups are then used for DOA estimation to generate multiple candidate angle sets, where each set contains a unique true solution and multiple pseudo-solutions. Finally, the cross-distance minimization method is applied to the multiple candidate angle sets to select the corresponding true solution for each group, and the final DOA estimation is given by combining the multiple true solutions. Simulation results show that as the number of antennas tends to large-scale, the proposed method can rapidly find the true solution for each group and achieve excellent estimation performance.