The paper introduces the HealMe model, a novel approach to psychotherapy using Large Language Models (LLMs). HealMe focuses on cognitive reframing, a key technique in cognitive-behavioral therapy (CBT), to help clients detach from negative thoughts and perspectives. Unlike previous LLMs that primarily convert negative emotions to positive ones, HealMe emphasizes fostering self-discovery and rational thinking. The model employs empathetic dialogue based on psychotherapeutic frameworks, guiding clients through distinguishing situations from feelings, brainstorming alternative viewpoints, and developing actionable suggestions. HealMe is evaluated using comprehensive psychological metrics designed specifically for assessing cognitive reframing in AI-simulated and real-world therapeutic conversations. Experimental results show that HealMe outperforms other models in terms of empathy, guidance, and logical coherence, demonstrating its effectiveness and potential positive impact on psychotherapy. The model's ability to maintain high empathy and guidance, along with its structured cognitive reframing process, makes it a promising tool for mental health applications. However, the study also highlights limitations, such as the model's inability to handle complex issues beyond three rounds of dialogue, and future work is needed to address these challenges.The paper introduces the HealMe model, a novel approach to psychotherapy using Large Language Models (LLMs). HealMe focuses on cognitive reframing, a key technique in cognitive-behavioral therapy (CBT), to help clients detach from negative thoughts and perspectives. Unlike previous LLMs that primarily convert negative emotions to positive ones, HealMe emphasizes fostering self-discovery and rational thinking. The model employs empathetic dialogue based on psychotherapeutic frameworks, guiding clients through distinguishing situations from feelings, brainstorming alternative viewpoints, and developing actionable suggestions. HealMe is evaluated using comprehensive psychological metrics designed specifically for assessing cognitive reframing in AI-simulated and real-world therapeutic conversations. Experimental results show that HealMe outperforms other models in terms of empathy, guidance, and logical coherence, demonstrating its effectiveness and potential positive impact on psychotherapy. The model's ability to maintain high empathy and guidance, along with its structured cognitive reframing process, makes it a promising tool for mental health applications. However, the study also highlights limitations, such as the model's inability to handle complex issues beyond three rounds of dialogue, and future work is needed to address these challenges.