From 1977 to 1983, non-uniform probability creel surveys were conducted on Lake Norman, North Carolina, and Lake Wylie, North Carolina and South Carolina, to estimate seasonal and annual creel parameters of pressure, success, harvest, and angler preference. Despite their geographical proximity on the Catawba River, mean annual pressure and harvest estimates on Lake Wylie were >3 times higher than respective estimates on Lake Norman. Catch rates for the 2 reservoirs were similar. Based on harvest, the Lake Norman sport fishery was a crappie and largemouth bass fishery, while the Lake Wylie fishery was a catfish, largemouth bass, and crappie fishery. The substantial productivity difference for these 2 edaphically similar reservoirs was attributed to certain physical/chemical characteristics of the reservoirs. Lake Wylie's much larger drainage area, smaller surface area, shallower basin morphometry, and shorter retention time resulted in higher annual nutrient loading and subsequent potential productivity. Sport fish harvest predictive models (morphoedaphic index and chlorophyll a index) were applied to both reservoirs and found to provide predictions within the range of natural variability. The author concludes that reservoir geographical proximity does not necessarily mean resource similarity. Consequently, broad management practices on a regional basis may not be optimal management policy. For proper resource management and utilization, reservoirs should be considered on a case by case basis. When manpower and budget constraints prohibit this, however, predictive models can be useful in considering reservoir differences based on physical and chemical parameter differences.