Summary Reservoir models are often built by matching the seismic reflection and production data individually. History matching is a highly non-linear inverse problem where by perturbing the model parameters we try to match the fluid flow dynamic response of a given Earth model with the observed production data. Geostatistical seismic inversion is a geophysical inverse problem that tries to infer subsurface elastic models that produce synthetic seismic data that matches the recorded seismic data. In spite of their different physical principles, both of these inverse problems have the same parameter and solution space. We propose in this work a simultaneous geostatistical iterative inversion methodology, where the retrieved subsurface models match simultaneously the observed seismic reflection and the reservoir production data. This methodology is based on stochastic sequential simulation as the model perturbation technique and uses a genetic algorithm in order to converge the inverted models into the solution from iteration to iteration. The proposed methodology was successfully tested and implemented in a challenging synthetic dataset, where the inverted petro-elastic models match considerably well the real models. In addition to the match between petro-elastic models, the production profiles resulting from the inverted models agree with the historic production data.