This paper presents a new interactive evolutionary approach for the vehicle routing problem. After a short review of main interactive optimization systems, we describe an evolutionary algorithm based on a set of human-computer interaction techniques. With a graphical display of the current population of solutions, a human can visually identify interesting route segments and tag them accordingly. Human annotations are taken into consideration in the evolutionary operators during the optimization process. Our interactive optimization approach is under experimental validation and user tests on benchmark problems are carried out to investigate the human-algorithm interaction.