This thesis aims to examine if there is any difference of the anaerobic threshold when Bruce protocol, increasing speed and grade stepwisely; Balke protocol, increasing grade only ; and Ellestad protocol, loading speed only are performed on the subjects. The subuects are 9 students of the physical education department at K university(20-21 years old). The experiment was performed from July 5 to July 21, 1999(3 weeks), at K university science laboratory. To examine the difference of the anaerobic threshold between the three treadmill protocols, One-way repeated ANOVA was performed (p＜.05), and Turkey method was used for post-hoc. The result of the anaerobic threshold`s difference between the treadmill protocols was found statistically significant (p＜.05). As a result of the post-analysis of the anaerobic threshold`s difference between the treadmill protocols, there was a statistically significant difference between Bruce and Ellestad, (p＜.05); However, there was no statistically significant difference between Bruce and Balke. The result can be explained two ways. First, the exercise load method of increasing grade gives the human body more burden. Second, there are chances of AT`s deferment because the subjects, students of the physical education department, who are accustomed to speeding, can run efficiently. Therefore, if the AT levels between the treadmill protocols are the same, the athletes, whose maximal oxygen uptake is large, and the non-athletes can use the same exercise load method. But as the AT levels of the treadmills protocols are different, the methods should be diverse according to the objects, exercise level, and the exercise type. For example, track runners on the field should use Ellestad to elevate threshold levels. Balke protocol is recommended for runners of height training to make them accustomed to grade load. And the runners like marathoners who are training or the condition that grades are changed need to choose Bruce protocol, because the Bruce protocol is heavily loaded.