In operation of a large LNG processing plant there will be uncertainties related to planning and realization of an optimal operation strategy. Results show the importance of including the whole production chain in the optimization. For a test scenario where we look at the event of delayed ship arrival, this gives lower losses than a normal approach where the upstream part of the system and the LNG process plant are optimized individually. The system is modeled by simple models of each main component starting at the wells and near-well region and ending with the export tanks for LNG, LPG and condensate. These proxy models can be regarded as first order approximations of the real system. The individual models are nonlinear, some are static and others are dynamic models. The models have been coarsely validated using a high-fidelity simulator of the value chain at the Snøhvit LNG plant. Use of these quite simple models combined with model-based optimization offers and interesting and feasible approach to optimize production in case of various events that require readjustment of the production planning.