Intelligent manufacture of electrical equipment has emerged as a key objective in the field of industrial production as a result of the exponential growth of both information technology and artificial intelligence. Traditional finite element analysis methods, however, generally need substantial numerical computations, resulting in sluggish computing speed and trouble reaching the requisite precision, and so failing to fulfill the real-time requirements, when applied to large-sized and complex-structured electrical equipment. In this post, we use deep learning to analyze motors’ electromagnetic fields. This work integrates neural networks with the task at hand, specifically by combining Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM). CNN excels at extracting key features from the input model, while LSTM is capable of learning long-term dependencies as a time-sequential network. By combining the advantages of these two models, the CNN-LSTM model is employed. To enhance prediction accuracy further, improvements are made to the CNN-LSTM model, resulting in the proposal of a prediction model named TCNA-LSTM. TCN is utilized to acquire long-range information, effectively reducing redundant features and providing a larger storage capacity. Additionally, residual connections facilitate the transmission of shallow information to deep layers, while attention mechanisms capture crucial information in the training data.