Abstract: Water conservancy and hydropower projects have posed a great threat to the fish habitat environment and migration, even the gene exchange between fish populations, where the river connectivity is restructured via manmade canals. Fishways are an increasingly popular solution to facilitate the passage of fish migrating. However, the flow speed of fishway depends mainly on fish swimming ability. Consequently, the body length of fish is a key factor in fish swimming ability. In this study, a series of stepped velocity tests were carried out to evaluate the induced flow speed (Uind), critical swimming speed (Ucrit), and burst speed (Uburst) of target species using the modified Brett-type swimming respirometer. A new fishway was also developed in Hainan Province of China to explore the effect of fish body length on swimming ability and flow speed. An Origin 9.0 software was used to transform three fish speeds and body lengths for linear regressions. Kaplan-Meier curves (Survival curves) were plotted using cumulative percentages and speeds in fish swimming ability tests. The suggestions were also given on flow speeds for fishway design. The results indicated that the relatively induced flow speed, critical swimming speed, and burst speed (bl/s) decreased with body length increasing. Linear relationships were achieved to describe the fish speeds and body lengths. The fitting efficiency of linear equation (R2 and F) after natural logarithmic data transformation was better than that without transformation, and the R2 values increased from 0.664-0.725 to 0.907-0.954. The effective speed for 90% of tested fish (EF90) was set as the flow speed in the vertical slot fishway. The EF90 value for Uind (90% induction rate, Uind-EF90) was 0.21 m/s, while the values for Ucrit and Uburst (90% success rates, Ucrit-EF90 and Uburst-EF90) were 0.59 m/s and 0.74 m/s, respectively. According to the standard design of vertical slot fishway realized by the Guideline for fishway in water conservancy, and Design code for fish passage facilities in hydropower projects in China, the flows of a fishway were recommended as follows: 1) Entrance. The flush water should be provided if the flow speed of the entrance was lower than fish-induced flow speed. The recommended entrance speed was 0.35-0.47 m/s (0.6-0.8 Ucrit-EF90) if the Ucirt-EF90 was 0.59 m/s, where the critical swimming speed of 0.6-0.8 m/s was generally set for attracting fish to the entrance. (2) Pool. The recommended pool speed was 0.21-0.59 m/s (Uind-EF90-Ucrit-EF90). The reason was that the pool speed was lower than burst speed, where the fish cannot lose their orientation. Since the fish critical swimming speed was the maximum sustainable swimming speed, the fish can commonly swim several hundred meters of fishway (100-500 pools), where the fishway design considered fish swimming endurance. 3) Vertical slot. The recommended velocity of the vertical slot was 0.57-0.74 m/s (minimum Uburst-Uburst-EF90). The reason was that the flow speed of fishway should lower the maximum swimming speed of fish, where the fish commonly used burst swimming to pass vertical slots. 4) Exit. The recommended exit speed was 0.21-0.50 m/s (Uind-EF90-0.50 m/s). The exit flow should be smooth and guide fish upstream, exceeding the induced flow speed of <0.50 m/s. These demonstrated that the fish speed decreased relatively, as the body lengths increased using natural logarithmic function in flow speed of fishway design.