Existing emotional support conversation (ESC) systems often react directly to surface-level distress expressions, generating reassurance or advice without adequately modeling the underlying psychological needs driving the distress. However, effective emotional support requires an interpretive process that connects what is expressed with why it is expressed and ultimately what support is needed. We argue that current ESC systems lack such an interpretive layer between observation and supportive intervention. To address this gap, we propose HierINT, a psychologically grounded hierarchical interpretive schema for emotional support conversations. HierINT distinguishes observable communicative intentions from the underlying psychological interpretations that drive them, including cognitive appraisals and unmet fundamental psychological needs. Building on this schema, we construct HierINT-Dlg, a dataset with hierarchical interpretive annotations and need-aligned support strategies. We further propose HierINT-CoRe, a collaborative reasoning framework that operationalizes HierINT by jointly modeling psychological interpretation and support strategy planning prior to response generation. Experiments on ESConv and HierINT-Dlg demonstrate consistent improvements over surface-level and intention-aware baselines, particularly on empowerment and personalization metrics. Our results suggest that effective emotional support requires moving beyond reactive response generation toward psychologically grounded need-informed intervention. Our data and code are available at https://github.com/43zxj/HierINT.

In emotional support conversations, unclear intentions can lead supporters to employ inappropriate strategies, inadvertently imposing their expectations or solutions on the seeker. Clearly defined intentions are essential for guiding both the supporter’s motivations and the overall emotional support process. In this paper, we propose the Intention-centered Emotional Support Conversation (IntentionESC) framework, which defines the possible intentions of supporters in emotional support conversations, identifies key emotional state aspects for inferring these intentions, and maps them to appropriate support strategies. While Large Language Models (LLMs) excel in text generating, they fundamentally operate as probabilistic models trained on extensive datasets, lacking a true understanding of human thought processes and intentions. To address this limitation, we introduce the Intention CEntric Chain-of-Thought (ICECoT) mechanism. ICECoT enables LLMs to mimic human reasoning by analyzing emotional states, inferring intentions, and selecting suitable support strategies, thereby generating more effective emotional support responses. To train the model with ICECoT and integrate expert knowledge, we design an automated annotation pipeline that produces high-quality training data. Furthermore, we develop a comprehensive evaluation scheme to assess emotional support efficacy and conduct extensive experiments to validate our framework. Our data and code will be publically released to facilitate further research.

Understanding the reason for emotional support response is crucial for establishing connections between users and emotional support dialogue systems. Previous works mostly focus on generating better responses but ignore interpretability, which is extremely important for constructing reliable dialogue systems. To empower the system with better interpretability, we propose an emotional support response generation scheme, named Emotion-Focused and Strategy-Driven Chain-of-Thought (ESCoT), mimicking the process of identifying, understanding, and regulating emotions. Specially, we construct a new dataset with ESCoT in two steps: (1) Dialogue Generation where we first generate diverse conversation situations, then enhance dialogue generation using richer emotional support strategies based on these situations; (2) Chain Supplement where we focus on supplementing selected dialogues with elements such as emotion, stimuli, appraisal, and strategy reason, forming the manually verified chains. Additionally, we further develop a model to generate dialogue responses with better interpretability. We also conduct extensive experiments and human evaluations to validate the effectiveness of the proposed ESCoT and generated dialogue responses. Our dataset, code, and model will be released.