Abstract Estimation of individual dry matter intake (DMI) of pastured beef cattle is critical to grassland management, genetic improvement of efficiency, and increasing the sustainability of animal agriculture. Currently, the only way to measure individual DMI of grazing cattle is through expensive, labor intensive, and spatially demanding research protocol. Very few options for assessment of individual DMI of grazing animals exist for cattle producers. We have developed a machine learning approach to predicting individual DMI of pastured beef cattle that requires only body weight (BW), water intake, and open-source climate variables. Training data were collected in Wardensville, WV at the West Virginia University Research, Outreach, and Education Center. Crossbred steers (n = 32) were phenotyped for Residual Feed Intake (RFI) in the barn. Animals were split into four groups of two low RFI and two high RFI. After a grazing acclimation period, animal groups were rotationally grazed concurrently through 0.05 ha plots for 7 d. Daily BW and water intake data were collected throughout the experiment using an RFID-equipped front-end scale and metered waterers. Beginning on d 2, animals were provided a bolus of Cr2O3 (10 g) at 0700 and 1900 h. Beginning on d 5, fecal samples were collected from each animal at 0700 and 1900 h. Samples were analyzed for Cr2O3 content and used in conjunction with forage digestibility from paired clipped samples to generate individual intakes. These data were used as ground truth data to assess model performance. We used our Repeated Measures Random Forest (RMRF) model to predict individual DMI of animals using daily BW, water intake, and climate variables. Intakes estimated by RMRF were compared with intakes estimated by NASEM equations. Our model outperformed NASEM predictions by 70%, indicating that our method achieves better results than the current most efficient approach. Though iterative model improvements will be made (collection of more training data, collection of more diverse training data, and use of first-differenced, moving average, and seasonally adjusted variables), our work demonstrates a novel, efficient method for researchers and producers to more accurately estimate DMI of individual pastured beef cattle.