Emotional self-awareness, the ability to recognize and articulate one's own feelings, is essential for emotional regulation and adaptive coping. However, many individuals struggle to connect with their inner states, particularly in East Asian contexts such as China, where emotional restraint is culturally emphasized. To address this challenge, we developed a large language model (LLM)-based chatbot intervention designed to enhance emotional self-awareness by guiding individuals to identify and reflect on their feelings regarding a personally distressing interpersonal event. In our preliminary study, participants who engaged with the emotion-focused chatbot showed greater improvements in emotion identification and description compared to those in the control condition discussing neutral lifestyle topics. Moreover, participants in the emotion-focused condition exhibited greater decrease in state anxiety and stress immediately after the intervention compared to those in the control condition. These findings provide preliminary evidence that a structured LLM-based conversational intervention can effectively promote emotional awareness and improve momentary well-being.