Abstract: A data-driven research method for intelligent transfer and risk warning of active distribution networks is proposed to address issues such as changes in distribution network load characteristics caused by high proportion and large-scale access of distributed power sources, increased complexity of load transfer in distribution networks, and potential risks of heavy overload after transfer. Firstly, an initial transfer plan for the active distribution network is adaptively generated using the deep deterministic policy gradient algorithm. Next, a net load forecasting model is established using a combination of convolutional neural networks and long short-term memory networks, enabling precise net load prediction after transfer. Finally, based on the forecasted net load, an overload risk warning mechanism is developed to ensure the safe and stable operation of the active distribution network after power transfer.