In this study, the ability of adaptive neuro-fuzzy inference system (ANFIS) and response surface methodology (RSM) approaches for predicting the draft force of subsoiling tines was assessed. Results of ANFIS and RSM approaches were compared with the results of regression models, too. The draft force was evaluated as affected by the tines at three levels (subsoiler, paraplow, and bentleg), forward speed at four levels (1.8, 2.3, 2.9 and 3.5 km/h), depth at three levels (30, 40 and 50 cm) and wing width at two levels (with wing = 30 cm and no-wing = 0 cm) at four replications. Test results show that tine types, speed, depth, and wing width were significant on the draft force but quadruplet interaction effect of them. Moreover, the increment of forwarding speed, tillage depth and adding wing increased the draft force of all tine types. Field data were applied for the development of the regression, ANFIS and RSM models. The results of ANFIS part showed that Gaussian membership function (gaussmf) configuration was found to denote MSE of 0.0156 and R2 of 0.998, consequently, it was the best ANFIS model. The RSM and best regression models had a high correlation (R2 = 0.9927 and 0.9968, respectively), too while ANFIS model was the better than them to predict the draft force of subsoiling tines with higher accuracy. The RSM graphs showed the changes of the output variable (draft force) caused by changes of input variables (tine type, speed, depth and wing width) better than ANFIS graphs for their surfaces with higher pixels. Moreover, the optimization process for prediction of the draft force was obtained 4.22 kN for depth of 35.19 cm, the forward speed of 1.9 km/h and wing width of 26.97 cm using the RSM approach.