Abstract The pink shrimp was a very important fishery in the southern Gulf of Mexico, but it is currently collapsed. In the early 1970s, fishery yields peaked at about 27,000 t per year, of which the pink shrimp was 90%. At present, yields are around 3000 t per year, and pink shrimp contributes less than 1000 t of this. Several studies indicate the main cause is a continually decreasing recruitment rate over the last nearly 30 years, but particularly since the mid-1980s. Single species models cannot predict stock recovery: the best estimate indicates a recovery of almost 20% assuming a closure of 10 years and the current stable stock. An ecosystem approach, based on Ecopath with Ecosim software, is applied to explore harvesting strategies aimed at recovery of shrimp stock, and based on fleet control. To this end, an improved Ecopath mass-balanced model of the southern Gulf of Mexico ecosystem was used for simulations based on Ecosim. A time series of 25 years of biomass and fishing rate data for the pink shrimp, from an age-structured model, was used to calibrate previous model simulations. We also used monthly recruitment for the same period, and changes in vulnerability to predation as forcing factors. At this stage, the model fits independent data well. Harvesting strategies were simulated using different combinations of weighting factors for optimized criteria. These criteria included: net rent, jobs, mandate rebuilding shrimp stock, and conservation of ecosystem structure. Eighteen realistic outputs were obtained after simulations, but all of them resulted in similar conclusions (CV