Reservoir mudflats limit development of healthy fish assemblages due to the lack of structural habitat provided by plants. Seeding mudflats with agricultural plants may mimic floodplain wetlands once inundated and provide fish habitat. However, planting success is uncertain because of unpredictable water level fluctuations that affect plant growth. Decision support tools can quantify uncertainty that influences decision outcomes, thus reducing risk in the decision-making process. I used Bayesian Decision Networks and sensitivity analyses to quantify uncertainty surrounding mudflat plantings as supplemental fish habitat in four northwest Mississippi reservoirs. When averaged across all uncertainty, planting was the optimal decision only in Enid Lake. Response profiles identified specific contours within Enid, Sardis, and Grenada reservoirs at which planting was the optimal decision. No such contours were identified in Arkabutla Lake. These results provide a quantified basis for establishing best management practices and identifying key system states that influence decision outcomes.