In　text　information　processing　field,　with　the　rapid　development　of　information　technology　and　Internet,　it　was　no　longer　difficult　to　collect　large　scale　of　corpus　for　training.　Privacy　preserving　data　mining　was　needed　to　discover　accurate　patterns　without　precise　access　to　the　original　data,　especially　for　personal　privacy　information.　Firstly,　a　privacy　preserving　online　adaptive　spam　filtering　model　was　given.　Then　an　example　of　privacy　preserving　spam　filtering　based　on　hash　function　was　proposed.　Thereafter,　an　approach　of　partial　hiding　privacy　preserving　spam　filtering　was　discussed.　Theoretical　analyses　and　experiments　show　that　it　can　provide　better　privacy　in　online　spam　filtering.