In this study, drying characteristics of two improved parboiled varieties of rice (FARO 44 and FARO 52) at air temperatures of 36oC, 45oC and 50oC were investigated. The drying data were fitted to seven thin layer drying models, namely, Agbashlo et al., Henderson and Pabis, Logarithmic, Newton, Two Term, Verma et al., and Wang and Singh. The models performances were evaluated by comparing the coefficient of determination (R2), standard error (SE) and relationship between the experimental and predicted moisture ratios through nonlinear regression analysis. The main factor controlling the drying rate was temperature and falling rate period characterized the entire drying process. The moisture content of the parboiled rice samples was found to be in the range of 26.33-27.57% (wb) which reduced to 8.87-9.98% (wb) for FARO 44 and 29.37-30.27% (wb) which reduced to 14.33-14.98% (wb) for FARO 52 after drying for various temperatures of 36°C, 45°C and 50°C for eight hours. The R2 and SE varied between 0.9953 - 0.9997, 0.9917- 0.9998, 0.9991 - 0.9999 and 0.9907 - 0.9994, 0.9986 – 1.0000 0.9300- 1.0000 for FARO 44 and FARO 52 respectively for the seven models. The three best models at 36°C for FARO 44 was the Logarithmic followed by Modified Henderson, Pabis and Newton. At 45°C, the Two term model was the best followed by Verma et al., Modified Henderson and Pabis. While for FARO 52 drying behaviour at each drying temperature of 36°C, the best model was the Two term followed by Wang and Singh and Logarithmic. The Two term model was the best model at 50°C. The Two term drying models satisfactorily described the drying behaviour which produced randomized residual plots with highest R2 and lowest standard error of estimates and gave best fitting curves.