Durum and bread wheat are widely planted cereal crops that contribute immensely to global food security. To maintain and improve on crop yields, fertilizers are applied including nitrogenous fertilizers. However, there is limited research focusing on the effect of nitrogen application rate on observed and estimated durum and bread wheat yields in dryland environments. This study investigated the application of unmanned aerial vehicle (UAV) multispectral bands and vegetation indices using artificial neural networks (ANN) and multiple linear regression (MLR) models to estimate yields of durum and bread under different nitrogen fertilizer application rates. The ratio vegetation index (r = 0.29; P < 0.05) and normalized difference vegetation index (r = 0.26; P < 0.05) showed a low, but significant correlation with bread yield under 48 kg/ha nitrogen application. The ANN model outperformed MLR for yield prediction under all nitrogen rates and produced highest accuracy of R² = 0.7753, RMSE = 0.0825 t/ha under 24 kg/ha nitrogen application for durum. The key findings from this study highlight that UAV datasets and ANN models can be used to predict durum and bread yields in real-time which is beneficial for crop nutrient management. The methods from this study should be explored with more robust machine learning and larger datasets for optimal crop yield estimation.