An integrated watershed management framework that combines economic, hydrologic and GIS modeling is developed to study cost effective land retirement in multiple watersheds to achieve off-site sediment reduction goal. This integrated framework examines two alternative standards-a uniform standard under which each watershed is required to achieve the same sediment reduction goal and a non-uniform standard under which marginal cost of sediment abatement is equal across watersheds. Furthermore, for each standard, costs of abatement under two alternative rental instruments based on marginal cost of sediment abatement ($/ton) and uniform payments per acre ($/acre) are examined. Then the cost effectiveness of the four policy options (uniform standard with $/ton and $/acre instrument, non-uniform standard with $/ton and $/acre instrument) is discussed. The integrated framework is applied to 12 agricultural watersheds in Illinois Conservation Reserve Enhancement program (CREP) region. The watersheds varied in size between 29,995 and 70,849 acres. Cropland within 900 feet of streams-129,955 acres (33.4% of all cropland in the 12 watersheds)-is considered eligible for enrollment into the CREP. Consistent with Illinois' program, a sediment reduction goal of 20% is selected for all of the simulations. Policy implications from the empirical results are quite interesting. With either a $/ton or a $/acre instrument, the non-uniform standard, which equalizes marginal cost of abatement across watersheds, outperforms the uniform standard policy. With either a uniform or non-uniform standard, a $/ton instrument outperforms a $/acre instrument. The least preferred policy option, the uniform standard with a $/acre instrument, is 2.5 times as costly as the most preferred policy option, the non-uniform standard with a $/ton instrument. These results suggest that program administrators may want to consider a program that includes a non-uniform standard and a rental payment instrument based on marginal cost of abatement in order to achieve their objectives at least cost.