Due to wireless communication technologies, positioning technologies, and mobile computing develop quickly, mobile services are becoming practical and important on big spatiotemporal databases management. Mobile service users move only inside a spatial network, e.g. a road network. They often issue the K Nearest Neighbor (KNN) query to obtain data objects reachable through the road network. The challenge problem of mobile services is how to efficiently answer the data objects which user interest to the corresponding mobile users. Therefore, how to effectively modeling road networks, effectively indexing, and querying on the road networks have become a popular topic. Lu et. al. have proposed a road network model that captures the real-life road networks better than previous models. Then, based on their model, they have proposed a RNG (Road Network Grid) index for speeding up the KNN query on real-life road networks. The RNG index structure is a quad-tree structure and a point-based structure. However, in their model, they divide the double track road which U-turn is allowed at some parts. This modeling does not capture the real-life road networks accurately. Since they divide the road, this makes the number of points of the graph increase. The number of times of partitioning the graph increases. It increases the execution time of constructing the index structure. The format of the leaf node of the RNG index makes the search time increase. Moreover, the query processing on the RNG index structure has to visit the root repeatedly. This condition makes the search time increase. Therefore, in this thesis, we propose a network model that captures the real-life road networks. We do not have to divide the real-life roads when we map the real-life roads into graph. We map the real-life road networks into graph directly. Then, based on our network model, we propose an EBNA (Edge-Based Nine-Area tree) index structure to make the search time of obtaining the interest edge information quickly. The EBNA index structure is an edge-based index structure. We store all of the edge information on the leaf node. We can obtain the edge information directly. Each edge information entry has a pointer point to link edges. Links of each edge entry consist a graph. This graph makes the KNN query processing visit the root only one time. From our simulation result, we show that the performance of constructing the EBNA index is better than constructing the RNG index and the performance of the KNN query processing by using EBNA index is better than the KNN query processing by using RNG index.