The treadmill is often used in health; physiological, biomechanical and athletic training related research. Despite its convenience in speed and slope control, it may be one of the major error sources. Among the researches that used treadmill, only a few had discussed the reliability of the apparatus. The present study used two different types of treadmill (Trackmaster, TM500; Star Trac, 2000) to examine the speed and slope adjustment accuracy. The goniometer, automatic timer, and a highspeed camera (120 Hz) were used to measure the accuracy objectively. One healthy male subject (178 cm, 85 kg) was tested on three different speeds and two different slopes. The results indicated that: 1. The Trackmaster can control the average speed error within 2%, whereas the error of slope was up to 5%. This means if slope is the test protocol, calibration should be taken before hand. 2. The Star Trac (model 2000) had an average speed range from 1.61% to 6.23% without load and 2.44% to 4.36% with load. The error of slope was as high as 20%. The results indicate a general-purpose treadmill can ' t fulfill the demand of an experiment. 3. The Trackmaster demonstrate better adjustment from the instant speed results, but the 30m/sec2 instant acceleration could cause a shearing force of 5000 N. Whether this is a potential injury sources needs further investigation.