Purpose The minimal and robust clinically important difference (MCID and/or RCID) are essential in assessing the clinical significance of multidimensional fatigue inventory-20 and checklist of individual strength-fatigue subscale questionnaires changes scores. This is the first study to determine the MCID and RCID of these questionnaires in chronic stroke survivors. Materials and methods A total of 125 participants in an observational cohort study completed MFI-20 and CIS-fs before and after receiving multidisciplinary rehabilitation (cognitive behavioral therapy, graded exercise and adaptive pacing therapy). Anchor-based MCIDs and RCIDs were calculated using the mean change, the mean difference and the receiver operating characteristics methods. To evaluate the accordance between of distribution-based MCIDs (1 SD, ½ SD, SEM, 1.96 SEM and MDC values) with anchored values, the accuracy, sensitivity, specificity and Youden’s index were calculated. Results The anchored MCIDs were between −5 to −7.33 for MFI-20 and −4.87 to −5.40 for CIS-fs. The anchored RCIDs ranged from −5 to −13.88 and −6 to −9.88 for MFI-20 and CIS-fs, respectively. The values of ½ SD and SEM for CIS-fs were consistent with anchored RCIDs. Conclusions The estimated MCIDs and RCIDs of MFI-20 and CIS-fs can help researchers and clinicians interpret their chronic stroke patient data.