The invention discloses a recommendation method based on user interest association, which mainly comprises the following steps of: establishing a user interest model; calculating the correlation among different users to the same interest point based on the user interest model; based on the user interest model, calculating the correlation of all interest points of the users; calculating related interest point recommended values of other users possessing of the same interest points under the current interest points of the users; and ranking the recommended values and recommending a plurality of interest points ranked the top. According to the recommendation method, the correlation among the interest points is fully applied, and the optimal results in the local are recommended to the users.