In this research, we present a novel clustering-based prediction model designed to forecast new infectious disease cases in India. The model integrates unsupervised clustering techniques with time series forecasting methods to enhance the accuracy and reliability of predictions. By clustering regions with similar historical disease trends and environmental factors, our approach enables more localized and precise forecasting. We employed k-means clustering to group Indian states based on demographic, climatic, and historical disease incidence data. Following the clustering, we applied an ensemble of time series models, including ARIMA and Prophet, to predict future cases within each cluster. The model’s performance was evaluated using real-world data on infectious diseases such as dengue, malaria, and COVID-19. Our results demonstrate that the clustering-based prediction approach outperforms traditional state-level forecasting methods, particularly in regions with significant variability in disease incidence. The model’s adaptability to various infectious diseases and its capacity to provide region-specific forecasts make it a valuable tool for public health authorities in planning and resource allocation.