Suicide remains a critical public health concern, often stemming from the profound emotional distress caused by mental health disorders. In today’s digital age, social media platforms have become common spaces for individuals to express suicidal thoughts, largely due to the sense of anonymity they provide. Detecting suicidal ideation on these platforms is crucial for early intervention and for preventing this ongoing mental health crisis. This study introduces a computational approach for identifying suicidal thoughts by analyzing the emotional tone of text at the sentence level. Our method leverages a Long Short-Term Memory (LSTM) model to process emotional data and track fluctuations in emotion throughout the text. The results demonstrate that our approach can effectively detect subtle emotional shifts, even in the absence of explicit death-related language.