Compared to conventional automobiles, electric vehicle technology is thought to be a very promising alternative. Optimizing range and energy efficiency during the design process is made possible by an understanding of and ability to control the energy consumption of electric vehicles. The advent of artificial intelligence algorithms has enabled the prediction of energy consumption in electric vehicles by leveraging their technical specifications. This research focuses on modeling and predicting the combined electrical energy consumption of electric vehicles through a novel methodology based on the Kolmogorov-Arnold network. This innovative approach greatly improved forecast accuracy and computing efficiency, which produced mean absolute error (MAE) of 0.7912 on the training dataset and 0.7766 on the testing dataset. Similarly, the R2 value for training and testing was 0.9712 and 0.9715, respectively. Furthermore, the performance metrics derived from the proposed symbolic formula closely matched those produced by the Kolmogorov-Arnold Network model.