This study aimed to determine how the pelvic algorithm affects the functional aspect of walking ability in chronic stroke patients based on the concept of NCR. Four patients with chronic stroke hemiplegia were included in this study. Functional training was conducted five times a week, 30 minutes per session for 3 weeks, using pelvic algorithms based on the NCR concept. The pelvic algorithm evaluated the hip joint and lumbar motility and displacement of the sacroiliac joint that affected gait function after stroke and provided an intervention method accordingly. The following measurement tools were used: the 10-meter walking test, Berg balance scale, Timed up and go test, and modified Barthel index. The measurements were performed before intervention, and 1, 2, and 3 weeks after intervention. The modified Barthel index was implemented 1 month before and after arbitration. After 3 weeks of intervention, the participants’ 10-meter walking test score increased by an average of 7.9 m/sec, showing an improvement rate of 22.8%; the Timed up and go score increased by an average of; the Berg balance scale score increased by an average of nine points, showing an improvement rate of 41.8%; and the modified Barthel Index increased by an average of five points, showing an average improvement rate of 9%. These results indicated that the application of pelvic algorithms based on the NCR concept helped to improve balance and gait in chronic stroke patients.