We present a fluid-flow constrained inversion approach for integrating controlled-source electromagnetic data and production data. In this approach, we assumed that the reservoir model has been well defined from a priori knowledge obtained from other independent measurements such as seismic and/or well-logs. Our objective was to reconstruct the permeability distribution and the shape and location of the flooding front. A finite-difference reservoir simulator was used to model the water-flooding process to simulate the time-dependent production data as well as the temporal and spatial distributions of water saturation and salt concentration, which were then transformed into the reservoir resistivity distribution using a petrophysical relationship. A finite-difference frequency-domain electromagnetic forward-modeling code was then employed to simulate the controlled-source electromagnetic response. The permeability distribution was reconstructed using a multiplicative-regularized Gauss-Newton algorithm for jointly inverting controlled-source electromagnetic and production data. From the water saturation distribution, we can identify the shape and location of the fluid-front resulting from the recovery process.