Stress in part of life affecting people throughout the world and if not dealt with it results to health complications. There are important reasons to point to the need for early detection and prevention of poor consequences and effects of stress. This work introduces a dual analysis model which analyses stress indices extracted from EEG data in the temporal and frequency domain. To achieve high accurateness and consistency of the tension recognition model, machine learning approaches are employed. The goal line is to improve the accuracy and reliability of stress recognition process, so that intervention can be promptly provided. Giving a brief introduction of stress, its effect on mental health, and the need for automated stress detection systems, the paper starts. The permutation of the EEG signals is emphasized to have a role of identifying stress related neural activities. Over the course of this study, different performance metrics including accuracy, sensitivity, specificity and F1 score are used to evaluate the model when it comes to detecting stress. As observed, the Hybrid Approach yields a higher accuracy than the current conventional approaches.