Biomass, catchability, and natural mortality are key parameters in fish stock assessment. Yet, it is difficult to estimate these quantities, especially natural mortality, when only fishery data are available. Using a method of population depletion analysis, we estimated these population and biological quantities for the white banana prawn (Penaeus merguiensis) in Australia's valuable Northern Prawn Fishery. In addition, we directly included fishing power change over time. The models were implemented in a Bayesian framework by incorporating process error, observation error, and random variability for the underlying parameters. The posterior median initial fishable biomass ranged from ∼2000 to 7000 t year−1, and the median catchability varied from ∼3.8 × 10−4 to 7.3 × 10−4 boat-day−1, resulting in an average fishing power increase of 2.6% per year. An unexpected result is the estimate of exponential natural mortality rate of ∼0.03 week−1. This value is substantially lower than an earlier estimate of 0.05 week−1, which was based on a single year's fishery data in one stock region and has been widely used for over four decades without validation. We attribute this low natural mortality estimate mainly to prawn aggregation behaviour.