Rare diseases have problems with low number of cases, low social awareness, and long time of diagnosis. “Targeted doctor” is the first step to help rare disease patients start the correct path of diagnosis and treatment. This article introduces the design of a decision-making engine for patients with rare diseases by constructing a knowledge graph of rare diseases and experts, using an intelligent question-and-answer system, and combining big data and artificial intelligence methods. This engine can perform rare disease pre-screening based on patient portraits and other information, and recommend the best visiting route to patients, thereby improving the efficiency of rare disease patients’ medical service system and enhancing the decision-making ability of rare diseases.