COVID-19 spreads across the world and specific pre-caution strategies are required for different regions depending on the current satiation. Therefore, proper region-specific pre-caution processes occupy a significant place to tackle with COVID-19 pandemic. During the COVID-19 pandemic, significant data are cumulated and these data can be utilized in order to cope with pandemic efficiently via providing a better understanding for decision-makers. In the aforementioned aspect, data related to the COVID-19 pandemic is used to decide on group states where the application of the same pre-caution processes has become efficient and effective. Therefore, COVID-19 indicators (e.g. number of deaths and infected) can be utilized to cluster the states, regions, countries, etc. In order to accomplish the underlined objective, data with seven features (rate of one dose, rate of two doses, number of cases, death, tests, recovered people, and percentage of positive tests) are retrieved for each of 50 states in the USA. After that, a dissimilarity matrix for cities is generated with respect to the corresponding seven features. Lastly, clustering methods (K-means, Agglomerative Hierarchical and BIRCH clustering, and P-median model) in literature are applied to gather clusters of states. In the proposed study, 50 states are taken into consideration and four different methods are applied to divide states into 6 subsets. The best result is gathered via the K-means application.