The "patch model" of optimal foraging theory (OFT) assumes that animals choose and exploit patches of food in order to maximise their food intake rate and derives the foraging rules (patch selection and exploitation) to achieve this. While evidence from field studies has been found to be broadly consistent with the predictions of OFT, this does not imply that animals actually forage using OFT rules, however, as the same observations could possibly arise as the result of non-optimising foraging rules (e.g., avoidance of low food intake rate). Because OFT is potentially very useful for the purpose of managing animal production systems (e.g., for predicting and manipulating the intake of grazing dairy cows), it is of interest to know to what extent OFT remains useful if animals actually behave suboptimally. This paper examines this question by simulating animals foraging in a two-dimensional spatially heterogeneous environment (New Zealand A07039; Online publication date 13 March 2008 Received 24 May 2007 accepted 19 December 2007 dairy cows grazing grass pasture) using six different foraging rules with varying degrees of optimality. The results showed that both optimal patch selection and optimal patch residence time rules significantly increase intake rate in this type of environment, but differences resulting from use of different variations of the patch selection rules are small. Therefore, although OFT is highly simplified, and animals' senses and knowledge are imperfect, this need not preclude the use of OFT in practical applications, as the predictions of OFT are likely to be similar to the predictions of other, more realistic, foraging strategies, even when these are quite sub-optimal in formulation.