Depression is a prevalent mental health disorder characterized by persistent sadness, loss of interest in activities, and a range of physical and emotional problems, affecting daily functioning and quality of life. Its complexity and the variability of symptoms across individuals pose significant challenges for accurate and timely diagnosis, necessitating advanced and nuanced detection methods. The paper presents a novel approach to depression detection by integrating an adaptive feature weighting mechanism into the Extra Trees Classifier, a machine learning algorithm under the ensemble learning category. By doing that, the approach is able to strengthen the model’s attention to important features, ultimately enhancing the model’s accuracy and efficiency for depression diagnosis. Firstly, the classifier is trained at the start to rate the features’ importance, and afterwards, it penalizes the dataset by these weights to fine tune the model which will appropriately rank informative features more. The method was demonstrated by using the pre-processed EEG dataset, and some quality outcomes that have been made with the aid of the weighted Extra Trees Classifier as a proof of the potential usefulness of the algorithm in mental health diagnosis. The significance of this research is to show that the capacity of the Classifier to effectively treat big datasets and ultimately generate positive outcome results, hence, creating an avenue for the utilization of machine learning in early diagnosis and management of depression.