Purpose:　Traditional　psychological　crisis　intervention　requires　one-to-one　and　long-term　consultation　by　psychotherapists,　which　has　high　time　cost　and　expense.　The　public　has　a　strong　realistic　demand　for　more　convenient　psychological　crisis　intervention　services.　In　order　to　solve　the　above　problem,　this　paper　presents　an　approach　which　transforms　the　actual　professional　psychological　consultation　data　into　knowledge　graph,　establishes　a　terminology/ontology　of　psychological　crisis　intervention,　and　constructs　an　automatic　service　prototype　system　of　psychological　crisis　intervention.　Method:　This　paper　takes　actual　professional　psychological　crisis　intervention　cases　as　prior　knowledge,　expresses　and　stores　them　in　the　form　of　triples.　The　natural　language　generation　model　based　on　GPT-2　model　is　constructed　by　using　the　data　in　PsyQA　as　the　training　set　for　several　rounds　of　training.　The　terminology　of　psychological　crisis　intervention　is　constructed.　Then　the　psychological　crisis　intervention　system　is　constructed,　including　modules　as　keyword　extraction,　deep　semantic　matching,　knowledge　graph　retrieval　crisis　intervention　strategy　providing　module　and　natural　language　generation.　Result:　The　test　results　show　that　for　the　users　which　can　correctly　input　the　psychological　problem　description　text,　the　system　can　Identify　crisis　types,　provide　the　corresponding　intervention　strategies　and　return　the　targeted　crisis　intervention　text.　Conclusion:　This　system　avoids　the　constraints　of　traditional　crisis　intervention　such　as　high　price,　poor　timeliness　and　few　professionals,　and　provides　a　feasible　method　for　people　who　need　to　get　simple　crisis　intervention　services　in　time.　©　2022,　The　Author(s),　under　exclusive　license　to　Springer　Nature　Switzerland　AG.