In recent years, electric vehicles (EVs) have gained popularity due to their low energy consumption and simple structure. However, the issue of limited driving range has not been effectively addressed. In order to alleviate the “range anxiety” phenomenon of drivers, this paper proposes a hybrid neural network model based on gated recurrent unit (GRU) and deep neural network (DNN) to accurately predict the energy consumption of EVs. The driving data is divided into time series and non-time series data based on their characteristics. It then employs GRU network to train the time series data, integrating the fitted values with non-time series data within a DNN network for further learning. Processing time series and non-time series data separately reduces computational burden and improves model performance by allowing feature engineering to focus more on the unique properties of each data type. The proposed hybrid neural network makes full use of the advantages of different network structures in learning characteristics, more comprehensively captures various influencing factors of electric vehicle energy consumption, and improves the learning ability of the model. The comparative results illustrate that GRU-DNN can obtain higher prediction accuracy, and the convergence speed is better than the comparison algorithm.