In the work robust hub location problem under demand uncertainty is presented. The proposed concept is based on serving the most likely and penalized demand. The scenario-based stochastic programming formulation of described problem is presented, where the objective function consists of the hub setup cost and transportation cost, the losses for unserved demand. The unserved demand penalty term in objective function is formulated as weighted expectation of absolute deviation between served demand and scenario demand. The Bender’s decomposition algorithm in application to introduced problem is shown.