Cognitive radio is the enabling technology for license-exempt access to the TV White Spaces (TVWS). There is ever increasing demand of users in the broadcasting and communication services. Large portions of unused spectrum in the UHF/VHF bands exist in India which can be used on geographical basis. This paper describes a study on path loss variation in UHF/VHF bands in India. The aim of this study is to develop and optimize a path loss model based on Linear minimum mean square error estimation (LMMSE) for India. We propose the LMMSE based Optimized Perez-Vega Zamanillo propagation path loss model. The measured path loss values, collected across India, are compared with proposed Optimized Perez-Vega Zamanillo path loss model and other existing path loss models. It is found that Optimized Perez-Vega Zamanillo propagation path loss model has the least root mean square Error (RMSE) of 13.98 dB. Other existing path loss models have root mean square Error(RMSE) value greater than 24 dB. Therefore, Optimized Perez-Vega Zamanillo propagation path loss model is best suited for predicting coverage area, interference analysis in India for TVWS.