To ensure timely and efficient roadside equipment maintenance during typhoon conditions, this study proposes an intelligent personnel scheduling system that integrates deep neural network (DNN) assistance with a branch-and-bound method. The scheduling problem is formulated as an integer programming model, incorporating constraints such as extreme weather impact, personnel availability, and emergency response prioritization. By learning from historical maintenance scheduling data, the proposed DNN-assisted branch-and-bound method optimizes the decision-making process for branch selection and pruning. Experimental results demonstrate the method’s effectiveness in generating high-quality schedules under various typhoon impact scenarios.