A computer simulation model was used to optimise strategies for marker assisted selection with best linear unbiased prediction (MBLUP) for quantitative traits in a selective breeding program for the southern Australian abalone industry. The simulation model assumed five marker loci associated with genes affecting growth rate explained 50% of the genetic variance for this trait. Use of markers associated with improved disease resistance was also modelled. Parameters and simulation scheme used are described. The simulations accounted for the possibility of early induction of reproductive condition to reduce generation intervals and for selective genotyping. Affects on profitability were calculated using estimates from the industry of income and expenses for a typical farm and accounting for the costs associated with marker testing. While the extra genetic gains from MAS were in the order of 15%, if the cost of genotyping each marker per animal was $10 AUS, then MBLUP for growth was only profitable when 75 progeny per family were genotyped. Early selection (halving the generation interval) was advantageous for MBLUP and BLUP provided the number of progeny per family was greater 5, as the advantage of two generations of selection more than compensated for the decreased accuracy of selection for early growth. However accurate estimates of parameters, such as the genetic correlation between early growth rate and growth and age at first spawning, would need to be obtained before considering implementing this strategy. The advantage of MBLUP was greater for disease resistance than for growth rate. Selective genotyping, which would greatly reduced genotyping costs of implementing MBLUP to improve disease resistance, did not greatly reduce the additional response achieved with marker assisted selection.