A new bio-economic deterministic approach and two crossbreeding simulation models i.e., the Stable model (ST) and the Variable model (VR) for 10 years of overlapping generations were developed. In the ST, the flocks` size and system is stable and in the VR these may vary, as necessary. There are up to 3 flocks of 2 sire and dam breeds. A computer program examines all combinations to find the optimal system and flocks` structure applying cash flow discounting. Each model proposes one or other of the rotational-, rota-terminal-, two way- and terminal-crossing systems under a detailed year-by-year approach, diagrammatically. The models were compared using the data attributed to the Australian Merino and Border Leicester. The ST required more ewes at the start and resulted in a higher cumulative net profit per ewe. It had less variation in the meat quality by use of generation preference and better utilization of breed effects due to less variation in the meatier breed`s gene contribution and to the stable flocks` size. The VR needed more ewes to be as profitable as the ST in the end. Nonetheless, with same initial numbers of ewes for both models, the VR totally gained more profit due to a larger cumulative number of the ewes raised. Further, in the VR, the meat quality was unstable and there was too much delay in gaining a notable annual net profit. However, where there are low initial investments and limited resources available at the commencement of crossbreeding, the VR would be more suitable.