Depression, also known as depressive disorder, is a mental health condition characterized by a persistent loss of interest or prolonged sadness. It should not be confused with regular mood changes that occur in daily life. If left untreated, depression can severely impact family relationships, work performance, and social interactions. In extreme cases, individuals experiencing intense depressive episodes may endure unbearable emotional pain, which can tragically lead to suicidal thoughts or actions. This paper proposes a novel deep learning algorithm to identify the depressive state of individuals by analyzing the content they share on social media platforms. The proposed algorithm utilizes a four-layer neural network integrated with an adjustment function based on the emotional tone derived from textual analysis. Experimental results show that the algorithm achieves high accuracy when tested with sample data collected from various social media sources.