<p>The rapid growth of the electric vehicle (EV) sector necessitates a robust and secure infrastructure to meet customers’ growing demands. An efficient management system called the Electric Vehicle Charging Station Management System (EVCSMS) is necessary because the Internet of Things (IoT) ecosystem greatly adds to the enormous amount of data generated by EV charging stations. However, due to its growing reliance on IoT infrastructure, cyberattacks targeting it are frequent. This paper describes an Intrusion Detection System (IDS) based on deep learning that is intended to identify fraudulent activity in IoT networks that facilitate EV charging stations. The proposed methodology makes use of an authentic IoT dataset from the CICEVSE2024 dataset and applies thorough data preprocessing procedures, such as normalization and cleaning. Features from a Siamese convolutional neural network is extracted together with important statistical features like mean, standard deviation, kurtosis, and skewness. To improve the detection process, Improved Secretary Bird Optimization (ISBO) is used for feature selection. A Multi-Head Attention-based Shuffle Net (MHA-SN) is used in the detection phase to precisely identify possible cyber threats. The reliable operation of EV charging stations and the continuation of daily activities can be ensured by implementing this advanced IDS, which can significantly reduce the risk of cyberattacks. Customer satisfaction and trust are increased when EV charging stations are secure. The widespread use of electric vehicles depends on dependable and secure services. The proposed method is implemented in MATLAB, and the performance is assessed for metrics like accuracy, precision, recall, and F-measure.</p>