The Facility Site Selection model for emergency services stations seeks to determine the optimal location for new facility stations. This decision-making model presents the optimal solution and different choices to enable governmental institutions to select the best option. Previous studies focused on either maximizing the service coverage area or minimizing the travel distance. However, maximizing the service coverage area still leaves the problem of facility service station location unsolved. Conversely, exclusively minimizing the travel distance adds many facility stations, which is not a viable solution owing to budget constraints. Moreover, previous studies considered a limited number of datasets to determine the location of new facility stations. This study proposes a Site Selection model that solves the problem of coverage area within a specific distance by using a clustering-based approach to add new facility stations that will lead to an increased coverage rate of demand. Specifically, the Elbow method is used with K-Means clustering, which is adapted to accommodate the number of new facility stations (optimal k). Further, a broad range of parameters in this study collected local data from multiple resources to build a local dataset for our use case (Dubai city), including point of interest, population, incident location, road network, and the existing police stations. We compared our (Elbow method with K-Means clustering) against three clustering algorithms (KMedoid, MiniBatchKMeans, and Fuzzy C-Means). Moreover, we compare our model against three previous models regarding the algorithm (proposed model), parameters, evaluation, and results from analyzing their performance. The results indicate that our proposed model provides superior performance by considering existing facility stations with multifarious parameters to determine the optimal location for new facility stations. The proposed model was applied in a real case study (Dubai police stations) after adding (7) new police stations within a speci