Climate modelling and prediction studies play crucial role in identifying suitable mitigation techniques to minimize or avoid adverse consequences of climate extremes. The accurate spatially and temporally distributed temperature and rainfall dataset are key components in climate prediction studies. Reanalysis datasets provide better spatial and temporal coverage than observational datasets; therefore, reanalysis datasets are widely used for global and regional studies. However, before using the reanalysis dataset in climate modelling studies, it is crucial to compare the robustness and accuracy of the reanalysis dataset with the observational dataset. In this study, daily gridded maximum and minimum temperature datasets of Indian Meteorological Department (IMD) (1° × 1°) and Sheffield (0.25°×0.25°) are compared using 62- years data i.e 1951-2012. The comparison is based on differences in spatial distribution pattern, probability distribution functions plots and box-plots of the respective gridded dataset. The spatial distribution of grid-wise averaged maximum and minimum temperature dataset generally compare well across pan India in both IMD and Sheffield; however, the significant differences are observed over western Himalaya (WH) and northeast (NE) region. The probability distribution of the pooled mean minimum temperature dataset of IMD is found significantly different from Sheffield using the two-sample Kolmogorov-Smirnov (KS) test. This study will be helpful for researchers who are planning to use Sheffield gridded temperature dataset for climate modelling studies.