The invention discloses a personalized recommendation method and system. The method comprises the following steps that input data is obtained, wherein the input data comprises the click behavior data and the click object data of a user; the input data is preprocessed; a first searching model and a second searching engine so as to conduct interest analysis on the preprocessed input data, and a first candidate list and a second candidate list are obtained; the first candidate list and the second candidate list are set and analyzed respectively so as to obtain the algorithm weight of the first candidate list and the algorithm weight of the second candidate list, and weight fusion is conducted; the result obtained through the fusion is reordered so as to serve as a recommendation list, and then the recommendation list is output. By means of the personalized recommendation method and system, the semantic comprehension problem of an ambiguous text in a recommendation scene is effectively solved, the accuracy of personalized recommendation results is ensured, and the recall quantity of the recommendation results is increased.