This study was aimed at finding an optimum pre-treatment method and parameters for convective drying of pumpkin for preparation of flour to maximize the retention of nutrients, for the production of pumpkin flesh powder. Artificial Neural Network (ANN) was employed for modeling of drying conditions to obtain powders with desired properties. For this, the pumpkin (Cucurbita moschata) flesh was subjected to blanching (90ᵒ C for 5 minutes) and citric acid treatment (0.1% acid for 30 minutes). Then they were subjected to cabinet drying at temperature and drying time varied at five levels. ANN modeling was performed using three different training functions and at two levels of hidden layers. The physico-chemical analysis had shown that blanching treatment came out with best physical properties and higher retention of nutrients when compared to the control and acid treated samples. The values of co-efficient of determination (R2), Root mean square error (RMSE), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) were used to determine the training function and hidden layer combination for each response variable for their prediction with highest accuracy.