Very recently, large outbreak of COVID-19 cases all around the world has also whacked India since approximately 30,000 cases confirmed within first three months of transmission. The present study used long-term climatic records of air temperature (T), rainfall (R), actual evapotranspiration (AET), solar radiation (SR), specific humidity (SH), wind speed (WS) with topographic altitude (E) and population density (PD) at regional level to investigate the spatial association with number of COVID-19 infections (NI). Bivariate analysis failed to find any significant relation (except SR) with the number of infected cases within 36 provinces in India. Variable Importance of Projection (VIP) through Partial Least Square (PLS) technique signified higher importance of SR, T, R and AET. However, Generalized Additive Model (GAM) fitted with log-transformed value of input variables and applying spline smoothening to PD and E, significantly found high accuracy of prediction (R2= 0.89), thus, well explained the complex heterogeneity among association of regional parameters with COVID-19 cases in India. Our study suggests that comparatively hot and dry regions in lower altitude of the Indian territory are more prone to the infection by COVID-19 transmission.