The invention discloses a social activity recommendation algorithm in a mobile social network, comprising the following steps: S1, calculating label similarity between any user and any social activity; S2, calculating geographical position preference degree between any user and any social activity; S3, calculating friend similarity between any user and any social activity; S4, calculating user preference of any user on any social activity; S5, calculating similarity between any user and any friend of the user; S6, calculating collaborative-filtering-based preference between any user and any social activity; S7, calculating preference scores of users on social activities; S8, recommending one social activity to a target user according to the preference scores of the target user on all the social activities. By adopting the social activity recommendation algorithm disclosed by the invention, the limitation of mechanized recommendation results caused by content-based recommendation is eliminated, so that the recommended social activity has more pertinence and accuracy.