The Intergovernmental Panel on Climate Change Tier 2 (IPCC-Tier 2) guidelines provide 2 models; a comprehensive (IPCC-CMP) and a simplified (IPCC-SMP) model to predict DMI as obtaining actual feed intake measurements of livestock is challenging. The IPCC-CMP includes equations to calculate net energy requirements for body functions, which is then connected to DMI using digestible energy utilization efficiency (REM), and energy digestibility (DE). In the IPCC-SMP, DMI is simply a function of BW and DE. These models are yet to be evaluated systematically for prediction accuracy. The objective of the present study was to evaluate the IPCC-Tier 2 models and compare them to extant models such as Cornell Net Carbohydrate and Protein System (CNCPS) model and National Research Council-2001 (NRC) model to predict DMI using an independent data set. Two experiments using lactating Holstein cows provided 209 observations of DMI, milk yield, milk fat content, BW and DIM. The average values were 21 kg/d, 32 kg/d, 3.7%, 670 kg, and 188 d, respectively. The overall agreement between predictions and observed values were determined with the square root of mean square prediction error expressed as a percentage of average observed value (RMSPE). Systematic biases of predictions such as mean bias (MB) and slope bias were also estimated and expressed as a percentage of RMSPE. The CNCPS relying on fat corrected milk yield and BW more accurately predicted DMI (RMSPE = 14.1%) than NRC (RMSPE = 19.4%), IPCC-SMP (RMSPE = 16.9%), and IPCC-CMP (RMSPE = 23.4%). The CNCPS model had minor systematic bias (<10%), whereas IPCC-CMP had a large mean bias (56.3%) for DMI to be over-predicted. The calculated average net energy requirements (e. g., maintenance = 0.47 MJ/kg of metabolic BW, and lactation = 95.2 MJ/d) were in line with literature values indicating that perhaps representations of REM and DE in the IPCC-Tier 2 methodology need to be revised for dairy cows, at least in North America. Overall, the results demonstrated that DMI of dairy cows can be predicted successfully using information such as milk yield, milk fat content, and BW that are routinely available in dairy farms.