Fluidized bed reactor models are generally specific to a single flow regime resulting in ambiguities and discontinuities at the regime boundaries. In practice, only the bubbling, turbulent and fast fluidization regimes are of industrial significance for catalytic reactions. The turbulent fluidization regime is especially advantageous because of improved interphase mass transfer, resulting in improved selectivities and conversions. It is shown that some of the difficulties in modeling can be resolved by means of the probabilistic-averaging model, recently published by Thompson et al. (1999). This model interpolates between the Grace (1984) two-phase bubbling bed model at low velocities and single phase axially dispersed flow for fully established turbulent fluidization conditions, leading to improved predictions of conversion and selectivity for catalytic fluidized bed reactors operated at flow rates covering the full range between bubbling and fully turbulent fluidization. An analogous approach should be useful for beds operated at higher gas velocities as fast fluidization conditions are approached.