Abstract: Location based recommendation systems are used with many applications that support maps and user’s interested locations. Recommendations must be accurate and near to user’s needs, especially when many applications depend on maps and locations. This study studies the effect of distances between users and their location preferences, combined with our previous work 3D-SVDs algorithm. Here, Manhattan distance is used as a measurement’s feature between location blocks against user’s location, this technique is injected into our recommendation algorithm and evaluated by using information retrieval recall and precision for different distances. The results are differed according to location distance and the recommendation lists are varied between the accuracy and diversity.