Accurate estimation of standing crop and herbaceous fuel moisture content (FMC) are important for grazing management and wildfire preparedness. Destructive sampling techniques have been used to accurately estimate standing crop and FMC, but those techniques are laborious and time consuming. Existing nondestructive methods for estimating standing crop in tallgrass prairie have limitations, and few studies have examined nondestructive estimation techniques for FMC in this environment. Therefore, our objective was to develop robust models for nondestructive estimation of standing crop and FMC in tallgrass prairie. We calibrated and validated stepwise multiple linear regression (SMLR) and artificial neural network (ANN) models for standing crop and FMC using data collected in tallgrass prairies near Stillwater, Oklahoma. Day of year (DOY), canopy height (CH), Normalized Difference Vegetation Index (NDVI), and percent reflectance in five wavelength bands were candidate input variables for the models. The study spanned two growing seasons and nine patches located within three pastures under patch burn management, and the resulting data set with >3 000 observations was split randomly with 85% for model calibration and 15% withheld for validation. Standing crop ranged from 0 to 852 g m^-2, and FMC ranged from 0% to 204%. With DOY, CH, and NDVI as predictors, the SMLR model for standing crop produced a root mean squared error (RMSE) of 119 g m^-2 on the validation data, while the RMSE of the corresponding ANN model was 116g m^-2. With the same predictors, the SMLR model for FMC produced an RMSE of 26.7% compared with 23.8% for the corresponding ANN model. Thus, the ANN models provided better prediction accuracy but at the cost of added computational complexity. Given the large variability in the underlying datasets, the models developed here may prove useful for nondestructive estimation of standing crop and FMC in other similar grassland environments.