Objective: Access to modern laboratory measurement methods may not be possible for every time to examine the effect of exercise or similar intervention. However, the reliability of the preferred measurement methods in such cases is scientifically very important. The aim of this study is to determine the reliability of easy applicable non-laboratory methods which can be used for determining upper body strength. Material and Methods: Thirty-three healthy men were participated in the study voluntarily. In the study resistance exercises: bench press, shoulder press, biceps curl and triceps extension exercises were measured in terms of 1 repeat maximum load. Callisthenic exercises: push up, pull up, dip and crunch were measured in terms of maximum number of repetitions. Kor exercises: trunk flexion (TRF), trunk extension (TRE), right lateral flexion (RLF), left lateral flexion (LLF) were measured in terms of maximum time. Reliability was analyses by the test-retest method. Results: There was no significant difference in test-retest for all 12 exercises. For three tests (TRE, RLF and LLF) intraclass correlation coefficient (ICC): 0.80 to 0.89; for nine tests ICC: 0.90 to 0.99 were calculated. In addition, typical error of measurement and smallest worthwhile change values of crunch, TRE, TRF, RLF, and LLF used for measuring center strength were found to be higher compared to other exercises. Conclusion: Consequently, total of 12 tests measuring different mechanical skills in this study were found to be reliable. It is recommend to increase of the number of familiarization trainings in the tests used to measure especially the central strength, and to repeat the tests at least twice for future studies for high reliability.