Freight　volume　forecasting　is　significant　to　highway　web　plan.　Here,　Support　vector　regression　optimized　by　genetic　algorithm　(G-SVR)　is　proposed　to　forecast　freight　volume.　We　adopt　genetic　algorithm(GA)　to　seek　the　optimal　parameters　of　SVR　in　order　to　improve　the　efficiency　of　prediction.　The　data　of　freight　volume　in　a　certain　port　from　1998　to　2007　is　used　as　a　case　study.　The　experimental　results　indicate　that　the　proposed　G-SVR　model　has　higher　forecasting　accuracy　than　grey　model,　artificial　neural　network.