This paper presents a hybrid recommendation approach that is used for discovering potential information with multidimensional clustering in recommender systems. This facilitates to obtain user groups for improving effectiveness and diversity of recommendation. The proposed algorithm works in three phases. In first phase, user groups are collected in the form of user profile, which applied multidimensional clustering algorithm and stored in the database for future recommendation. In second phase, the appropriate clusters are chosen using pruning of clusters. In third phase, the recommendations are generated for target user with similarity measures and quality rating prediction. The performance of proposed approach is evaluated using a public movie dataset and compared with two representative recommendation algorithms. The experimental results demonstrate that our proposed approach performs superiorly and alleviates problems, such as cold-start and data sparsity in collaborative filtering recommendation.