The design of soil covers as part of the remediation initiatives for mine waste facilities is common practice and more and more engineers are faced with designing these facilities. The design issues associated with these soil covers are complex and requires a good understanding of unsaturated flow. More engineers are making use of numerical models to predict how cover designs would perform, and this has been necessitated by the high non-linearity associated with solving unsaturated flow problems. There are numerous numerical models available to the prospective users and they are all simplifications of reality using different assumptions. The users of these models often do not understand the limitations of the models but rather select a model based on ease of use, and that lends to increasingly erroneous designs being proposed for construction. This paper describes a comparative evaluation of a cover design using four different codes available today; SoilCover, SWIM, HYDRUS-2D and HELP. The paper does not propose to measure these codes up against each other and make statements about which is better than the other but rather illustrates how the assumptions in each code affect the outcome. The data set used for the comparative study is a well calibrated data set that that has been collected over a long period in time and that has been collected according to strict controls. The authors present the pitfalls of numerical modelling of the surface flux boundary conditions, as well as present guidelines towards appropriate use of these modelling tools.