A semi-empirical two-compartment, constant parameter model is used to predict airborne and surface dust concentrations. The model parameters are air in- and exfiltration, internal particle sources, surface deposition caused by settling, Brownian and turbulent diffusion and thermophoresis, track-in of dust particles and resuspension. Model predictions are calculated for some typical scenarios, and the soiling rate of a vertical surface is calculated for a range of friction velocities and electric field strengths. Model sensitivity is determined based on input parameter value distributions for a population of rooms estimated from published data. The predictions are sensitive to track-in and resuspension rates on which field data thus are needed.