The　rapid　development　of　Internet　information　technology　makes　the　problem　of　information　overload　become　more　and　more　serious,　and　recommendation　system　is　one　of　the　effective　ways　to　solve　this　problem　which　is　favored　by　people.　However,　for　the　massive　data　information,　the　recommended　algorithm　faces　the　bottleneck　problem　of　processing　speed　and　computing　resources,　so　this　paper　proposed　a　parallel　collaborative　filtering　recommendation　algorithm　based　on　Spark.　The　RLPSO　algorithm　is　used　to　optimize　the　clustering　factor　of　the　K-　means　clustering　algorithm　by　associating　users　with　similar　interests　into　a　cluster　and　using　the　recommended　algorithm　for　users　to　recommend　is　implemented　on　the　Spark　platform.　The　experimental　results　show　that　the　improved　algorithm　has　a　significant　improvement　in　the　prediction　accuracy,　and　has　a　higher　speedup　and　stability　compared　with　the　traditional　collaborative　filtering　recommendation　algorithm.