Communicable diseases remain a significant challenge for public health management. In particular, for resource scarce settings, it is important to understand the linkages and dynamics of multiple diseases that share common resources, space or time. We develop a framework, called Multiple Disease Management Framework (MDMF) based on machine learning approach for managing multiple diseases occurring in close space and time to identify locations that experience high disease burden rates. We use 8 water related disease incidence data in Punjab, Pakistan from year 2013 to 2019 to investigate interactions among hotspots of different diseases. However, the model is scalable and can be applied to any number of diseases. The hotspot analysis involves multi-level clustering and tagging of individual disease incidence streams that generates a distance based graph over a geographical area and is then integrated into a single stream in the framework to identify final sensitive locations called cluster alarms. The initial individual disease clustering yielded number of clusters as high as 24 clusters for each disease with up to 16 neighboring clusters of other diseases of similar sizes. The cluster tagging and multi-level clustering process was able to identify as low as 19 locations of cluster alarms across the whole province of 38 districts. The identification of high disease hotspots and their dynamics with the neighboring hotspots of multiple diseases allows identification of locations with higher need of related public health resources. This identification is very critical for national health agencies for optimal allocation of resources and devising an effective intervention programs.