A numerical bioenergetics simulation model based on the physiological processes affecting individual clams across a range of phenotypes describing a cohort has been developed and applied to the conditions in Great South Bay, New York. The clam population is relatively sensitive to food and to a lesser extent to temperature within this system. The timing of temperature and food in the spring, and more importantly in the fall, can increase population sensitivity beyond the effects of one factor operating alone. The effects of fishing on the stocks in proportion to the size structure present, and as directed fisheries on various size classes (littleneck, cherrystone, chowder) was simulated. Recruitment overfishing was responsible for the stock decline in the 1970s and 1980s, but the continued decline into the late 1990s and 2000s cannot be attributed to fishing alone. Recruit-per-adult declined after the mid 1990s. Modeled stock recovery times under constant environmental conditions are on order of 10–15 or more years depending on the exploitation rate. Under base conditions a proportional fishery that removes approximately 25% of the stock, or a littleneck fishery that removes approximately 37.5% of that size class annually would provide the best economic returns under constant average environmental conditions. Slightly less harvest would be desirable to avoid overfishing in years of less than optimal environmental conditions.