Parameter uncertainties of existing fish stock assessment models may lead to incorrect management recommendations on optimal fishing strategies especially in tropical fisheries. Therefore, it is advisable to consider the effect of uncertainties in the population parameters when formulating management recommendations. This .paper attempts to evaluate the parameter uncertainties of the Beverton and Holt yield per recruit model to estimate optimal fishing strategies. A 10% variation on deterministic estimates of population parameters of Amblygaster sirm off the Western coastal waters of Sri Lanka was used to demonstrate the proposed method. Different magnitudes of perturbations of the estimated values of natural mortality, asymptotic weight and growth coefficient were considered and optimal management parameters of the Beverton and Holt yield per recruit model were determined. Multiple regression models were also formulated to describe the relationships between the uncertain input parameters and the optimal management parameters.