Lightning casualties cause tremendous loss to life and property. However, very lately lightning has been considered as one of the major natural calamities which is now studied or monitored with proper instrumentation. The lightning characteristics over India have been studying by using daily data low resolution time series and monthly data high resolution monthly climatology. We have used ANN time series method (a neural network) to analyze the time series and defined which one will be the best predictor of lightning over India. The time series of lightning is output(dependent) and input (independent) are k-index, AOD, Cape etc. The Gaussian process regression, support vector machine, regression trees and linear regression defined the input variables. Which show approximately linear relation.