This paper proposes the Emotional Chain-of-Thought (ECoT), a plug-and-play prompting method that enhances the performance of Large Language Models (LLMs) in emotional generation tasks by aligning with human emotional intelligence guidelines. The method incorporates Goleman's Emotional Intelligence Theory, which includes five dimensions: Self-Awareness, Self-Regulation, Motivation, Empathy, and Social Skills. ECoT guides LLMs through a process of emotion identification, emotional reasoning, and emotional response generation. To assess the effectiveness of ECoT, the authors propose an automated evaluation method called Emotional Generation Score (EGS), which evaluates LLMs' responses based on multiple dimensions of human emotional intelligence. The EGS is validated by comparing it with human expert evaluations, demonstrating its reliability and consistency. Experimental results show that ECoT significantly improves the emotional generation capabilities of LLMs, enabling them to generate more appropriate and positive responses. The study also discusses the potential applications of emotional generation in areas such as emotional chat assistants, emotional rewriting, and emotional captioning. The authors highlight the importance of aligning LLMs with human emotional intelligence to ensure that generated responses are harmless and beneficial to users. The study also acknowledges the limitations of current approaches, including the subjective nature of emotional generation and the need for further research in this area. Overall, the paper contributes to the field of emotional intelligence by providing a new framework for evaluating and enhancing the emotional generation capabilities of LLMs.This paper proposes the Emotional Chain-of-Thought (ECoT), a plug-and-play prompting method that enhances the performance of Large Language Models (LLMs) in emotional generation tasks by aligning with human emotional intelligence guidelines. The method incorporates Goleman's Emotional Intelligence Theory, which includes five dimensions: Self-Awareness, Self-Regulation, Motivation, Empathy, and Social Skills. ECoT guides LLMs through a process of emotion identification, emotional reasoning, and emotional response generation. To assess the effectiveness of ECoT, the authors propose an automated evaluation method called Emotional Generation Score (EGS), which evaluates LLMs' responses based on multiple dimensions of human emotional intelligence. The EGS is validated by comparing it with human expert evaluations, demonstrating its reliability and consistency. Experimental results show that ECoT significantly improves the emotional generation capabilities of LLMs, enabling them to generate more appropriate and positive responses. The study also discusses the potential applications of emotional generation in areas such as emotional chat assistants, emotional rewriting, and emotional captioning. The authors highlight the importance of aligning LLMs with human emotional intelligence to ensure that generated responses are harmless and beneficial to users. The study also acknowledges the limitations of current approaches, including the subjective nature of emotional generation and the need for further research in this area. Overall, the paper contributes to the field of emotional intelligence by providing a new framework for evaluating and enhancing the emotional generation capabilities of LLMs.