Mobile–immobile models (MIM) have rarely been used for inverse simulation of measurements of variably saturated water flow and contaminant transport. We evaluated two MIM approaches with water transfer across the mobile and immobile regions either based on relative saturation (Se) differences, MIM(Se), or pressure head (h) differences, MIM(h), for inverse simulation of transient water flow and Br− transport in aggregated soil. Six undisturbed Ap soil columns (14.7-cm length and diameter) at wet, medium, and dry initial water contents were subjected to application of 0.005 L Br− solution of 1000 mg L−1 and subsequent irrigation of 1 cm h−1 for 3 h. Two similar irrigations were applied after 7 and 14 d. Measurements comprised pressure heads at depths of 2.8 and 12.8 cm, average soil column water contents, outflow, and effluent solute concentrations. This experimental information was used for simultaneous optimization of van Genuchten soil hydraulic parameters for mobile and immobile regions, the dispersivity, and the first-order rate coefficients for water and solute transfer. In total, eight parameters were estimated for MIM(Se) and 10 parameters for MIM(h). The inverse MIM approaches described the various hydraulic and transport data adequately. Physical nonequilibrium was more pronounced for wet and dry than for intermediate initial moisture, while initial moisture had no obvious effect on the total Br− lost. For wet and dry initial conditions, parameter estimates seemed fairly reliable, with the exception of the highly correlated saturated water contents in mobile and immobile regions. MIM(h) yielded parameters that appeared physically more consistent with observations, but required smaller time steps than MIM(Se) to overcome oscillations of pressure heads. Both MIM approaches were found to be suitable for inverse simulation of physical nonequilibrium transport during variably saturated flow.