Emotion recognition in text plays a crucial role in enhancing the effectiveness of chatbots by enabling them to better understand and respond to users’ feelings. This work focuses on using a popular machine learning algorithm called logistic regression to categorize emotions from textual inputs. Through the use of labeled data with examples of different emotional expressions, the chatbot is able to predict the user’s emotional state during interactions with accuracy. The logistic regression model evaluates the text input to assign a probability distribution over various emotions by utilizing features like word frequency, sentiment analysis, and linguistic patterns. This makes it possible for the chatbot to adjust its responses appropriately, resulting in more sympathetic and customized exchanges. Through this approach, the chatbot can exhibit a higher level of emotional intelligence, thereby improving user satisfaction and overall user experience.