As internet communication has gone through a boom with social media, humans have been more and more using text to express their emotions, so it has become crucial to come up with ways to accurately capture these emotions beyond mere positive or negative sentiment analysis; this research proposes a novel method employing deep learning and semantic text analysis (DLSTA) to detect precise emotions such as joy, sadness, anger, fear, disgust, and surprise in written language with high accuracy by taking advantage of word embeddings to capture the meaning of words and their context better than other methods available. The paper discusses the challenges of processing large-scale unstructured internet data and emphasizes the need for advanced NLP techniques to improve emotion detection accuracy.