Caching frequently accessed data on mobile client is an important technique to improve availability of data and reduce access time. Cache replacement policies find victims from Mobile cache to get removed due to limited cache size. Traditional cache replacement policies depend on temporal locality of user’s access pattern. But these policies do not serve all the issues in mobile clients. As mobile client move from one location to other it exhibits temporal as well as spatial localities. Traditionally cache replacement policies consider temporal parameters like access pattern, frequency, probability, data size etc. For mobile environment there are many cache replacement policies were researched. In this paper we consider PAID (Probability Area Inverse Distance) policy and modified cost function of it to improve that policy. Previously Access probability, Area of Valid scope and Data distance these parameters are used to calculate cost of data item in cache. We modify the function and add new parameter Data Size. At the time of pre-fetching items for mobile cache, its associated data items are also get fetched. Due to which now mobile cache will have data which is associated with each other and spatial query will become local query. Test results shows that Modified Probability Area Inverse Distance (M-PAID) consistently improves performance of cache replacement in mobile environment.