This paper describes a deterministic age-structured dynamic pool model for floodplain-river fisheries. The model is applied to a heavily exploited floodplain fishery in north west Bangladesh to quantify the effects of hydrological modification and exploitation on production (catch per unit area) inside a typical flood control scheme, and to explore mitigating management interventions. Existing modifications to the hydrological regime were predicted to diminish production only marginally (10%), and similar differences in production may result from differences in the seasonal pattern of exploitation. Closing the fishery for a month was predicted to increase production by at least 30%, with greatest increases (up to 115%) predicted for closures during those months when fishing mortality is at its highest. The practicability of fishing strategies that maximise production are discussed. The model also predicts that diminished production caused by modifying flood season water heights inside flood control schemes may be compensated by increasing dry season water levels. Manipulating water levels in this way by means of sluice gates must, however, take account of the often conflicting needs of other sectors such as agriculture, and the importance of dynamic edge effects for maintaining the natural fertility of the floodplain. Sensitivity analysis indicates that more reliable model predictions may be achieved with more precise estimates of the parameters which affect recruitment.