In 1996, the NASEM beef cattle committee developed and published an equation to estimate cow feed intake using results from studies conducted or published between 1979 and 1993. The same equation was recommended for use in the most recent version of this publication in 2016. The equation utilizes body weight (BW), diet digestibility and milk yield. Our objective was to validate the accuracy of the NASEM equation using recently published and unpublished data. Criteria for inclusion in the validation data set included projects conducted or published since 2002, direct measurement of forage intake, adequate protein supply, and pen feeding (no tie stall or metabolism crate data). The validation data set included 48 treatment means for nonlactating cows and 29 treatment means for lactating cows. Quantitative data collected for the nonlactating data set included BW (593 ± 78.1 kg), BCS (5.7 ± 0.73), and Mcal NEm per kg of feed (1.26 ± 0.16 Mcal/kg) and lactating data set included DMI (12.7 ± 2.98 kg) and BW (505 ± 62.4 kg), BCS (4.6 ± 0.44), NEm per kg feed (1.25 ± 0.24 Mcal), and DMI (14.3 ± 2.08 kg), respectively. Non-intercept models were used to determine slope and bias when predicted DMI was regressed against observed DMI. The slope for linear bias in the NASEM nonlactating equation differed from 1 (P = < 0.0001) with a 13.9 percent downward bias. Similarly, when the NASEM equation was used to predict DMI in lactating cows, the slope differed from 1 (P < 0.0001) with a downward bias of 16.6 percent. Therefore, new prediction models were developed for both nonlactating and lactating cows. Log and exponential transformations were used to correct for heteroskedasticity. The best-fit nonlactating and lactating equations seem to provide need to be further validated with independent data sets. The current NASEM equation for predicting intake underestimated feed intake for both gestating and lactating beef cows. The new equation may improve the accuracy of predicting cow feed or forage consumption.