The use and uptake of models for freshwater and ecohydrology purposes has often been hindered by criticisms, including: poor match between model predictions and field observations; high variability or uncertainty of outputs; and the models being too complex and difficult to understand or to parameterise and implement. We argue that the gap between the development and use of such models can be closed through a better understanding of the role of modelling and through better modelling choices that better align to management needs, and problem and project contexts. We propose a framework that guides the ‘fit-for-purpose’ design, development and evaluation of models. The elements of the framework focus on three elements: useful, reliable, and feasible. This entails making modelling choices, including the selection of scales, system features and processes to represent and model type, that ensure the model addresses end-user and management needs, achieves an appropriate level of certainty and trust, and meets the practical constraints of the project. Models have become an important tool for determining environmental flow needs and assessing ecological outcomes of altered flow regimes in river systems. With the use of case studies drawn from the Murray-Darling Basin and basins in Northern Australia, we will demonstrate how the fit-for-purpose framework can facilitate the development of models that more effectively bridge science and management, and thus better inform adaptive flow management in practice.