Abstract: To address the issue of voltage limit violations at photovoltaic grid connection points, a multi-port energy router is designed. A reconfigurable distributed model predictive control (DMPC) strategy is developed based on different grid-connected point voltages and energy storage states of charge, to coordinate and control the output of each port of the energy router, achieving flexible adjustment and optimization of the voltages at photovoltaic grid-connected points. The proposed strategy consists of three layers of control. The upper layer coordinates the control of each port’s output based on the grid-connected point voltages. According to the day-ahead photovoltaic power output curve, the middle layer optimizes the energy storage energy control in advance. The lower layer is designed to smoothly shed the load or curtail PV to deal with certain extreme scenarios. The effectiveness of the proposed three-layer control strategy is verified through the construction of an energy router experimental platform and the design of three operational scenarios. The experimental results confirm the feasibility of the control strategy. Under varying operational scenarios, the energy router is capable of controlling each port to stabilize the grid-connected point voltage. This can effectively suppress voltage fluctuations and prevent voltage limit violations at the grid-connected points, thus ensuring smooth system operation when the status of each port changes.