Moving Horizon Estimation (MHE) is an efficient optimization-based strategy for state estimation. Despite the attractiveness of this method, its application in industrial settings has been rather limited. This has been mainly due to the difficulty to solve, in real-time, the associated dynamic optimization problems. In this work, a fast MHE algorithm able to overcome this bottleneck is proposed. The framework exploits the advantages of simultaneous collocation- based formulations and makes use of large-scale nonlinear programming algorithms and sensitivity concepts. The approach is demonstrated on a full-scale polymer process, where accurate state estimates are obtained and on-line calculation times are reduced dramatically.