The invention relates to the field of internet communication, and discloses a user group correlation degree-based personalized recommendation method, which comprises the following steps of: A, clustering users by using a clustering algorithm; B, judging distance from a target user to a cluster edge, executing a step C when the distance is greater than a given threshold value, otherwise executing B-1, calculating correlation degree between a cluster of the target user and other clusters, B-2, combining previous r clusters related with the cluster of the user, B-3, searching n closest neighbors in the combined cluster, and further executing a step D; C, searching n closest neighbors in the cluster of the target user; D, predicting a grading value of a related product according to grade of the closest neighbor on the product; and E, selecting the previous m products to be recommended to the user according to the level of a predicted grading value. The invention also discloses a user group correlation degree-based personalized recommendation system. According to the method and the system, the accuracy for personalized recommendation can be effectively improved.