Rapid and reliable techniques for estimating herbage mass (HM) are essential for germplasm evaluation and proper grazing management. Since hand‐clipping small quadrats is time consuming, mathematical relationships between easy to use, nondestructive techniques and hand‐clipped estimates of HM (double sampling) have been utilized. This study was conducted to quantify relationships between HM and three commonly utilized nondestructive sward measurement techniques [plant height (HT), settled disk meter height (DISK), and capacitance (CAP)] in field plantings of ‘Brazos’ bermudagrass [Cynodon dactylon (L.) Pers.]. Estimates of HM, HT, DISK, and CAP were obtained from the same location in irrigated plots and nonirrigated, grazed pastures. Linear equations were observed to describe the relationship between HM and HT [ Y = 414 + 12.9X (r2 = 0.86), Y = 37 + 21.7X (r2 = 0.86)] and CAP [ Y = ‐1276 + 2.4X(r2 = 0.77), Y = 1289 + 2.8X (r2 = 0.89)] for plots and pastures, respectively. The relationship between HM and DISK was linear [ Y = 282 + 29.3X (r2 = 0.91)] and quadratic [ Y = 421 + 44.9X ‐ 0.7X2 (r2 = 0.80)] for the pasture and plot studies, respectively. For all HM relationships, the ability to predict HM was generally better in pastures than in small plots. Problems associated with lodged herbage in plots greatly influenced the HM predictive relationships of all techniques presumably by increasing the variability at high HM levels. Due to the large differences in the range of HM, in tiller height and tiller density, separate equations were required to predict HM in pastures and plots, regardless of the technique utilized. No single technique was observed to best predict HM. In the pastures, CAP and DISK were more efficient at estimating HM, while in the plots HT was the most efficient estimator of HM. This study suggests, while double‐sampling may allow for the estimation of HM in bermudagrass, factors such as tiller height and tiller density will influence the predictive relationships of all techniques requiring separate prediction equations to be developed where differences in sward composition are expected.