Abstract: With the deep integration of renewable energy sources such as wind and photovoltaic power into integrated energy systems, user-side grid-connected integrated energy stations—acting as new bidirectional power users with both generation and load characteristics—introduce significant uncertainties. These uncertainties pose new challenges to the siting and sizing of distributed generation in active distribution networks, as well as to the safe, stable, and low-carbon operation of the grid. Therefore, precise modeling for the low-carbon siting and sizing optimization of distributed generation and energy storage is required. First, the electricity purchasing and selling behaviors of bidirectional power users with integrated energy systems under multiple typical scenarios are incorporated into the planning layer, enabling the effective transfer of user-side uncertainties into planning decisions. At the same time, the dynamic carbon potential calculation method based on carbon emission flow theory is applied to the distributed generation siting and sizing scenario of distribution networks with bidirectional power users with integrated energy systems, achieving precise assessment of carbon emissions under different scenarios. On this basis, a bi-level multi-objective optimization model is established, with the siting and sizing of distributed photovoltaic generation and energy storage as decision variables, and with comprehensive consideration of economic efficiency, carbon reduction, and operational safety. The model is solved using a multi-objective particle swarm optimization algorithm, and the technique for order preference by similarity to ideal solution (TOPSIS) is employed to select optimal planning schemes under different preference settings. Simulation results show that different planning preferences significantly influence the optimal configuration of distributed generation. The proposed method achieves coordinated optimization of planning and operation under multiple conflicting objectives and uncertainties, providing effective decision support for the low-carbon planning of active distribution networks with bidirectional power users.