Natural drying characteristics of sesame seeds (SS) were investigated under indoor conditions with both forced convection (FC) and natural convection (NC) of air. The drying kinetics of SS was characterized in terms of effective diffusion coefficient and resistance to diffusion. For the FC method, seeds were dried at a constant air velocity of 1.1 m s-1 and air temperature and relative humidity in the range of 25-29°C and 35-40%, respec- tively. For the NC method, air temperature and relative humidity were in the range of 32-36 and 30-35%, respectively. Six thin-layer drying models, namely, Khazaei, Peleg, Page, Handerson and Pabis, logarithmic, and Weibull, were fitted to drying data. Modelling the correlation between moisture ratio with drying time and drying method was also carried out using artificial neural networks (ANN). SS of average initial moisture content of around 50.8% (d.b.) were dried to the final moisture content of about 3.0-3.7% (d.b.) until no further changes in their mass were observed. The drying of sesame seeds took place in the falling rate period. During the FC experiments, the time to reach the final moisture content of 3% was found to be 400 min. The same moisture content of sesame seeds was found to achieve its equilibrium moisture content (3.7%) after 900 min when seeds were dried using the NC method. Thus, the FC drying times were around 55% shorter than the NC drying times. In the FC and NC drying methods, the drying rates of sesame seeds at the very beginning times of drying were equal to 22.47 and 6.9 (g H2O kg-1 dry matter min-1), respectively. The effective water diffusion coefficients of SS under FC and NC conditions were 3.1×10-11 and 1.1×10-11 m2 s-1, respectively. Corresponding values for overall resistance to diffusion were 70.8×105 and 19.6×106 m2 s kg-1, respectively. The results showed that the Khazaei model gave better fit than the other five models. The Peleg and logarithmic models also had an acceptable accuracy in predicting the drying kinetics of SS. The ANN technology was shown to be a useful tool for predicting the moisture ratio of SS as a function of drying method and drying time. The optimal ANN model was found to be a 2-6-3-1 structure with hyperbolic tangent transfer function. This optimal model was capable of predicting the moisture ratio with R2 higher than 0.998, RMSE of less than 0.0192 and MRE about 2.63%. It was concluded that ANN represented SS drying characteristics better than the mathematical models.