Point-of-interest　(POI)　recommendations　aim　at　identifying　candidate　POIs　and　ranking　them　in　a　descent　order　according　to　the　probabilities　of　a　user　visiting　them.　The　paper　takes　the　scalability　of　information　and　user　personalization　into　consideration　to　improve　POI　recommendation　service,　and　proposes　a　personalized　POI　recommendation　method　based　on　user　check-in　behaviors　in　online　social　networks.　First,　the　user＇s　travel　experience　in　the　target　region　is　used　to　reduce　the　range　of　candidate　POIs.　At　last,　the　proposed　method　ranks　the　candidate　POIs　to　meet　the　user＇s　personalized　need　by　combining　the　user　preference,　attraction　of　a　POI　on　the　target　user,　and　social　recommendations　from　friends.　Experimental　results　show　that　the　proposed　method　is　feasible　and　effective.