The characteristics of motorbike discharge behaviors are very important for the analysis of the capacity of exclusive motorbike lanes. However, previous researches disregarded the influences of different separation types (i.e. road marking or physical separation) and did not illustrate the methods of measuring lane-width, and the numbers of sampling cycles as well. The findings in previous studies were inconsistent and impractical. This study collected field data using cameras at eight sites with lane-widths ranging from 1.7 to 3.4 meters in Taipei city, with more than 50 cycles at each site. Researchers recorded the horizontal positions of back-wheels of motorbikes passing the reference line, which was identified as a stop line, and determined the actual discharge-width by an α-percentage cutting rule. In the meantime, the numbers of discharged motorbikes were counted every 2 seconds. The findings of this study were summarized as follows. (1) The value of α was 0.05 because the corresponding actual discharge-width could be effective in estimating saturation flow rates. The actual discharge widths at these eight sites ranged from 1.3 to 3.4 meters. (2) The discharge rates became stable 10 seconds after green light turned on. It showed that the saturation flow rate exists. (3) It was found that the saturation flow rates clearly increased when the available lane widths widened. (4) Based on the available lane-width (W, meters) and the corresponding saturation flow rate (S, motorbikes/hr), a simple regression model was estimated as S = 6168.6 + 441.9W2. The coefficient of determination (R2) was 0.90 with all estimated parameters being statistically significant. Furthermore, this study estimated the capacities under various green intervals based on the estimated saturation flow rates. The results of this study are essential for setting up exclusive motorbike lanes in the future.