Chinese sturgeon (Acipenser sinensis) is a critically endangered species. A flume-type respirometer, with video, was used to conduct two consecutive stepped velocity tests at 10, 15, 20, and 25 °C. Extent of recovery was measured after the 60-min recovery period between trials, and the recovery ratio for critical swimming speed (Ucrit) averaged 91.88% across temperatures. Temperature (T) effects were determined by comparing Ucrit, oxygen consumption rate (MO2), and tail beat frequency (TBF) for each temperature. Results from the two trials were compared to determine the effect of exercise. The Ucrit occurring at 15 °C in both trials was significantly higher than that at 10 and 25 °C (p < 0.05). The Ucrit was plotted as a function of T and curve-fitting allowed calculation of the optimal swimming temperature 3.28 BL/s at 15.96 °C (trial 1) and 2.98 BL/s at 15.85 °C (trial 2). In trial 1, MO2 increased rapidly with U, but then declined sharply as swimming speed approached Ucrit. In trial 2, MO2 increased more slowly, but continuously, to Ucrit. TBF was directly proportional to U and the slope (dTBF/dU) for trial 2 was significantly lower than that for trial 1. The inverse slope (tail beats per body length, TB/BL) is a measure of swimming efficiency and the significant difference in slopes implies that the exercise training provided by trial 1 led to a significant increase in swimming efficiency in trial 2.