Depression is a common and often underdiagnosed mental health disorder, with serious consequences for individuals’ health and well-being. Social networks such as X provide access to vast amounts of textual data that can be utilized to detect symptoms of depression. In this project, we utilized a dataset of 232,074 annotated tweets to detect depression on X. Tweets were manually annotated by mental health experts and labeled as positive or negative for the presence of depression symptoms. We employed text preprocessing techniques including Word2vec, TF-IDF, Glove Embedding, and fast Text to transform tweets into vector representations. Subsequently, we used neural network models such as LSTM and CNN to learn from these vector representations and perform binary classification to predict the presence or absence of depression symptoms in tweets. Our models findings showed accuracy, in identifying depression on X platform which's a breakthrough that could help identify depression early and assist users in finding the right treatment resources easily and quickly. Moreover this initiative has the potential to help reduce the stigma surrounding health conditions by encouraging individuals to seek support for their symptoms. The methods utilized in this project could also be applied to identify mental health issues, on social media platforms which can play a significant role in enhancing the overall mental well being of society.