Mathematical models are central to ruminant nutrition. Feed intake (FI) is paramount in the performance of livestock and has been of interest when creating said models. The aim of the present study was to evaluate 4 models in their prediction of FI in dairy cows fed total mixed rations (TMR): BSM-Milk (BioSimetrics Ltd.) a dynamic mechanistic whole cow model, the FI equation part of the CNCPS (Fox et al., 2004), the equation include in NRC (2001) and that in Feed Into Milk (FiM, Thomas 2004). A trial was carried out at SRUC’s Dairy Research Centre, Scotland UK. Two contrasting TMR were offered to 2 groups of 40 5 Holstein Friesian cows. The diets were: forage (kg/kg/DM Grass silage 0.40, maize silage 0.23, crimped wheat 0.11, beans 0.25 and min 0.01) and concentrate based (Wholecrop 0.40, Megalac 0.02, whey 0.08, min 0.01 and a concentrate 0.50). Using electronic feeders (HOKO, Insentec, the Netherlands) individual FI were recorded daily. Details of the animals and the TMR were used as inputs. FI predictions obtained were compared with those obtained on-farm. To evaluate the predictions regression analysis, the limits of agreement (LoA) method and the concordance correlation coefficient (CCC) were used. All statistical analyses were carried out using R. The evaluated models predicted FI with different levels of success. Obtained R² values were: 0.78 BSM-Milk, 0.48 CNCPS, 0.42 NRC and 0.47 FiM. The CCC were 0.88 BSM-Milk, 0.58 CNCPS, 0.61 NRC and 0.53 FiM. The LoA showed that BSM-Milk predicted FI in average 0.19 higher than observed (limits −3.80 to 4.19), similarly CNCPS predicted FI 0.98 higher than observed (limits −5.06 to 7.03), NRC predictions were −0.41 lower than observed (limits −6.80 to 5.98) and predictions made with FiM were 4.19 higher than observed (limits 0.28 to 8.85). BSM-Milk was the model with the best performance when compared with the rest of the evaluated models. Future research should compare BSM-Milk predictions to those obtained with models that use a more mechanistic approach to FI prediction.