Rainfall is a seasonal period in India, especially in the delta region. The cyclone period cause more rainfall leads flood to affect human life and agriculture rescues more effectively to make destroy. To avoid this flooding attack in seasonal time, to analyses the weather series data analysis to take precautions. In previous cases, the non-relational feature dependencies reduce the forecasting accuracy. To resolve this problem, we propose a spatial rainfall-intensive rate using PSO optimized Dense net Convolution neural network. Initially, the Delta region weather dataset is preprocessed to marginalize the features. The rainfall rate was analyzed through spatial data using spatial rainfall intensive rate (SRIR). To analyze the entity relation hit rate was estimated through the Seasonal rainfall Hit forecasting rate (SRHFR). Through the feature weight, Adaptive PSO feature selection (APSO) is applied to select the spectral features and trained with a Dense net convoluted neural network (DnCNN) to forecast the result based on flooding risk by category. This proposed system effectively attains high-performance evaluation in weather forecasting to predict rainfall and flooding level. The result performance shows the prediction accuracy as well as precision rate compared to the other system.