Purpose/Objective(s): Management of respiratory motion is important for tumors in the thorax and upper abdomen. Respiratory patterns could change over time even under respiration control. Stability of the respiratory waveform not only contributes to stability of tumor motion, but also is necessary for gating and tumor-tracking methods. The purpose of this study was to evaluate the change of respiratory patterns by repeating under different respiratory conditions. Materials/Methods: Data were acquired from 20 patients who underwent SBRT between September 2017 and December 2018. Three clinically used conditions, the free breathing (FB) condition, orally instructed (OI) condition and continuous audio coaching (AC) condition, were employed. The patients breathed without any instruction in the FB condition. They were instructed to breathe regularly and shallowly in the OI condition, and to breathe following the continuous voice instruction for inhalation and exhalation in the AC condition. Abdominal wall motion was used as a surrogate during the acquisition of the patients' respiratory waveforms. The lowest point (exhalation) of the respiratory wave was examined as the index of the fluctuation of the baseline, and the distance between the lowest point and highest point (inhalation) of the respiratory wave as the index of the fluctuation of the amplitude, in addition to the respiratory cycle. Each breathing method was performed for one minute and was repeated ten times. The mean value of each 1-min trial was calculated in all indices, and the standard deviation (SD) value of each 1-min trial was calculated in the lowest point of the respiratory wave. We calculated the correlation between these data and the number of trials in each index. Results: Significant correlations were found on the lowest point in the FB condition (Spearman's rank correlation coefficient (rs)Z-0.709, pZ0.0275), the SD of the lowest point in the FB condition (rsZ-0.709, pZ0.0275), the distance between the lowest and highest points in the OI (rsZ-0.879, pZ0.00198) and AC (rsZ-0.721, pZ0.0242) conditions, and the respiratory cycle in the OI condition (rsZ-0.879, pZ0.00198). Single regression analysis showed a linear correlation on the lowest point in the FB condition (yZ0.117-0.00329x, R^2Z0.532, pZ0.0167), the SD of the lowest point in the FB condition (yZ0.0760 e 0.00250x, R^2Z0.529, pZ0.,0171), the distance between the lowest and highest points in the OI condition (yZ1.34 e 0.0162, R^2Z0.657, pZ0.00444), and the respiratory cycle in the OI condition (yZ3.94 e 0.0441x, R^2Z0.774, pZ0.000786). Conclusion: By repeating respiration, the fluctuation of the baseline decreased although the baseline declined in the FB condition, and the fluctuation of the amplitude decreased in the OI and AC conditions. Based on these findings, an individual approach to respiratory management should be considered.