Climate perturbations like changes in the frequency of heavy rainfall, droughts can directly affect the surface water quality and also there is some direct and indirect of effect of physical processes into it. Limitations in the sampling frequency, detection methods and prediction of changes with the climate changes, water monitoring and management programs often fail to protect the water pollution and consequently public health. Application of deep learning and machine learning algorithms with the emergence of spatial information technology can make it feasible to evaluate and predict the surface water quality. The aim of this research project is to develop a model-based decision support framework which focuses on source water quality monitoring by analyzing the spatial variability of extreme rainfall and to predict the changes in the surface water quality associated with it. A case of Toowoomba region of South East Queensland of Australia will be taken as this region went through extreme rainfall and flood events in the last decade. The water supply sources of Toowoomba Regional Council is affected due to the extreme rainfall but no study is conducted to analyse how the change of climate is affecting the surface water sources. Smart data driven approach for the prediction of water quality and management was not conducted so far. The proposed integrated approach will provide a centre of attention for engaging the utilities and consumers, can identify the priority location where change is required, entitle better understanding of service refinement and will ensemble confidence among policy makers, management bodies and consumers in water and catchment management sector.