In this paper, a hybrid genetic-VNS algorithm is proposed for the no-wait flowshop problem with total flowtime minimization. To avoid pitfalls of GA, such as poor local search capability and premature convergence, a rather effective VNS local search is introduced based on the framework of the improved GA. To fast convergence of the algorithm, ICH2 (an efficient composite heuristic) is used for the initial population generation. Experimental results show that the proposed algorithm outperforms other best two recent existing methods on both small and large instances.