Abstract: Online optimization method for aero-engine was mainly studied in this paper.Trust region filter algorithm was applied to solve aero-engine online optimization problem in which a relative nonlinear component-level model was used as the optimization model.The algorithm has better optimality and feasibility than the basic trust region algorithm(BTRA).Sub optimization problem infeasibility is escaped by using a restoration algorithm,and the filter algorithm guarantees the algorithm global convergence ability.As shown in a simulation example,the trust region filter algorithm was used in the minimum specific fuel consumption mode for some two-spool turbofan engine performance seeking control,and the simulation results show that it has better optimization effectiveness than BTRA,if it's intended to solve aero-engine online optimization problem.