Reliable air temperature forecasts are necessary for mitigating the effects of droughts and Heatwaves. The numerical weather prediction(NWP) model forecasts have significant biases associated and therefore need post-processing. Post-processing of temperature forecasts using probabilistic approaches are lacking in India. In this study, we post-process the Global Ensemble Forecast System (GEFS) and EuropeanCentre for Medium Range Weather Forecasts (ECMWF) NWP model temperature forecasts for short to medium range time scales (1-7 days)using two probabilistic techniques, namely, Bayesian model averaging(BMA) and Nonhomogeneous gaussian regression (NGR). The post-processing techniques are evaluated for temperature (maximum and minimum) predictions across the Indian region. Results show that the probabilistic approaches considerably enhance the temperature predictions across India except the Himalayan regions. These techniques also comprehensively outperform the traditional post-processing techniques such as the running mean and simple linear regression. The NGR performs better than the BMA across all regions and is able to provide highly skillful temperature forecasts at higher lead times as well. Further, the study also assesses the implication of probabilistic post-processing Tmax forecast towards forecast enhancement of heatwaves (HW) in India. Post-processed Tmax forecasts revealed that the NGR approach considerably enhanced the HW prediction skill in India, especially in the northwestern and central Indian regions, considered highly prone to HW. The findings of this study will be useful in developing enhanced HW early warning and prediction systems in India.