Objective To explore the perceptions of young and middle-aged patients experiencing a first-ever stroke regarding the use of DeepSeek to assist medical decision-making, and to provide a reference for healthcare professionals to help patients accurately understand the capabilities and limitations of artificial intelligence (AI) and to promote high-quality care decisions.Methods A descriptive qualitative design was used. Eighteen hospitalized young and middle-aged patients with first-ever stroke were purposively sampled between April and June 2025. Data were collected through semi-structured interviews and were analyzed using content analysis.Results Participants′perceptions of DeepSeek-assisted medical decision-making were summarized into three themes:readily accessible “doctor”（breaking temporal and spatial constraints, matching symptom information, and offering concrete care-seeking suggestions）; imperfect tool（inability to actively and comprehensively collect information, lack of vivid or vividly presented information, concerns about information accuracy, and concerns about privacy breaches）; and lack of caring warmth（expectation of face-to-face communication and need for emotional support）. Conclusion Young and middle-aged patients with first-ever stroke showed generally positive attitudes toward DeepSeek-assisted medical decision-making, while also expressing concerns and expectations. It is recommended that healthcare professionals provide health education through diverse approaches to help patients accurately recognize the capabilities and boundaries of AI, learn to use AI tools appropriately, promote high-quality care decisions, and avoid delays in seeking care.