We provide a robust optimal hedging strategy in an incomplete market. This policy can protect the investor from parameter uncertainty. The investor aims to minimize a function of hedging error under the worst case scenario by means of solving a min-max robust optimization problem. We apply this methodology to the asset and liability management and employ an expected shortfall hedging criterionas our value function. The robust policy is more conservative than the naive one when the fund is facing solvency risk. The investor can benefit from the robust policy when the expected return is overestimated.