The integration of numerous distributed energy resources into distribution networks (DNs) can induce large voltage fluctuations and network loss. We introduce a collaborative active and reactive power optimization (CARPO) method for DNs and microgrids (MGs) to efficiently improve the voltage quality and mitigate network loss. First, the CARPO method and models for the DNs and MGs (DMs) are intended to reduce voltage deviations, minimize network loss, and improve the operation efficiency of the entire system. Second, to protect MGs, we aggregate privacy-preserving feasible operation regions of the active and reactive power outputs from distributed energy resources in MGs. A scaled-down MG equivalent model, which ensures high accuracy, is derived for optimal DN operation. Third, based on the equivalent projection theory, the optimal operation flow of DMs with non-iterative projection method is achieved to reduce the computational complexity. The DM model is decomposed into sub-models for the DM levels. The optimal solutions of the coordination variables are obtained for MG power scheduling. Finally, the proposed CARPO method is evaluated through simulation in a modified IEEE 33-bus DN. The results demonstrate that the proposed CARPO method can optimize the system operation and improve the economy of DMs.