With　the　development　of　public　security　information　system,　a　large　number　of　food　safety　cases　are　stored　in　the　database.　This　study　uses　the　HMM　model　and　the　Viterbi　algorithm　as　text　processing　technologies.　We　downloaded　3555　food　safety　criminal　judgments,　Chinese　word　segmentation　is　an　effective　way　to　deal　with　them.　We　can　find　the　high　frequency　words　after　processing　the　segmentation　results,　the　key　words　of　the　case　were　extracted,　such　as　the　name　and　carrier　of　harmful　food　additives,　the　crime　time　and　the　crime　place.　The　Apriori　algorithm　is　used　to　analyze　the　internal　rules　of　common　food　safety　cases,　which　has　certain　reference　significance　for　combating　and　preventing　food　safety　crimes.