This paper presents a coordination strategy for decentralized real-time optimization for an energy community. To benefit users and support grid operation, electric vehicles are managed to schedule for charging and discharging of both active and reactive power together with cooperating with heap pumps. The decentralized optimization problem is formulated as a nonlinear programming problem with the presented penalty parameters and then solved using the primal-dual interior-point method. The coordination scheme, based on AC optimal power flow approach, is used to minimize load shedding when a grid violation event is expected. A modified IEEE European Low Voltage Test Feeder is implemented in PowerFactory, and the optimization algorithm is developed in a Python environment that links together for co-simulation. Experiments are performed with actual data taken from local meteorological in Sweden to demonstrate the effectiveness of the proposed method. Results from the numerical simulations show that the proposed coordination scheme can maintain grid stability while minimizing the household electricity cost, prolonging the lifetime of the EV battery, and smoothing indoor temperature compared to the uncoordinated schemes.