Based　on　the　analysis　of　the　Hadoop　open　source　distributed　computing　platform　as　well　as　the　parallel　training　methods　for　the　BP　network,　for　the　disadvantage　of　time-consuming　when　using　large　amounts　of　texts　to　train　the　BP　network,　we　designed　a　BP　network　text　categorization　model　based　on　data　parallel　method　on　Hadoop　platform　using　the　MapReduce　programming　model.　The　model　uses　the　method　of　batch　training,　it　adjusts　the　network　weights　after　getting　the　accumulated　error　by　summing　every　sample　training　error　on　each　node,　and　the　categorization　of　text　is　done　in　parallel.　The　method　based　on　Hadoop　platform　improves　the　training　speed　of　BP　network　and　efficiency　of　text　categorization,　and　achieves　good　categorization　performance.　©　2013　IEEE.