Resting-state fMRI suffers from poor test-retest reliability because of between-subject and within-subject variability. Scan duration is one the main factors affects the reliability of rs-fMRI studies. We showed that under anesthesia, the time needed to optimize ICC of rs-fMRI metrics including ALFF, fALFF, functional connectivity, and ReHo is lower compare with awake state. The optimum scan duration that satisfies good reliability is 14-20 min and 8-17 min in awake and under anesthesia, respectively. Also, variability of ICCs is lower under anesthesia than in awake.