When regulations are species specific but the species are part of a multispecies fishery, studies show that harvest rates are correlated such that net revenues attributed to each species are also correlated. This correlation suggests that portfolio theory is well suited for multispecies fisheries that exhibit joint productive characteristics. This paper uses a portfolio approach to model the behavior of fishermen faced with multiple targeting options in a randomharvest fishery. The approach draws from the expected utility hypothesis and financial portfolio theory to predict optimal targeting strategies. The methodology is applied to the pelagic longline fleet operating in the U.S. Atlantic Ocean, Caribbean, and Gulf of Mexico. The model provides evidence that area closures aimed at reducing juvenile swordfish mortality will be more effective in certain regions. Efficient risk-return frontiers are also generated for use in predicting targeting behavior in lieu of a closure.