It is necessary to explore a new experimental protocol of which the experimental channel has similar flow field with fishway to study the swimming ability. Firstly, the critical swimming speed (101.01±20.86cm/s) and the burst swimming speed (196.94±21.80cm/s) of fish were acquired in swim chambers by velocity increment tests. Secondly, based on the critical swimming speed of fish and the designed flow speed at vertical slot of some Hydropower Station’s fishway (110.00cm/s), an experimental channel with vertical slots for testing volitional swimming ability of fish was built by setting different trapezoid barriers in the channel. After this, the swimming tests of Schizothorax oconnori to cross water velocity barriers were carried out with two different barrier layouts. For one layout, the channel has four steps in two gradients (the flow speed in velocity barrier of condition 1 was 101.55±14.87 cm/s, while condition 2 was 114.63±24.28 cm/s), for the other layout, the channel has only one step (the flow speed in the 160 cm long velocity barrier of condition 3 was 137.45±17.63cm/s). 93.33% fish in condition 3 cross the single step barrier successfully in a swimming speed of 209.43±21.76 cm/s; the swimming speeds(214.01±30.64cm/s) in all three conditions were not significantly different from burst swimming speed of fish (P>0.05). The swimming trajectories of fish and the corresponding flow field show that the time cost and path length for fish to swim upstream is closely related to the path fish choose. Fish reduce its time cost and path length to swim upstream by taking advantage of flow which is in the same direction with fish movement. This method and result of this research could provide reference for fish design, modification and evaluation.