Aim To investigate the effects of a 12-week split-belt treadmill walking (Sb-TW) practice using an error augmentation strategy on temporospatial gait asymmetries, dynamic balance, and locomotor capacity in adolescents with unilateral cerebral palsy (ULCP). Methods Fifty-two adolescents with ULCP (age: 10–16 years) were randomized into either the Sb-TW group (n = 26; underwent repeated Sb-TW practice, with exaggeration of the initial step-length asymmetry, three times/week, for 12 sequential weeks) or control group (n = 26; received equivalent dosages of traditional single-belt treadmill training). Step-length and swing-time asymmetries, directional (LoSdirectional) and overall (LoSoverall) limits of stability, and locomotor capacity [6-minute walk test (6-MWT), Timed Up and Down Stair test (TUDS), and 10-m Shuttle Run Test (10mSRT)] were assessed pre- and post-intervention. Results The Sb-TW group demonstrated more favorable changes in step-length asymmetry (p < .001, η 2 partial = 0.27), LoSdirectional [affected side direction (p = .033, η 2 partial = 0.09), forward direction (p = .004, η 2 partial = 0.16), and backward direction (p = .01, η 2 partial = 0.12)], and LoSoverall (p < .001, η 2 partial = 0.31) than the control group. Also, the Sb-TW group showed significantly higher locomotor capacity [6-MWT (p < .001, η 2 partial = 0.38), TUDS (p = .032, η 2 partial = 0.09), 10mSRT (p = .021, η 2 partial = 0.10)] as compared to the control group. Conclusion The Sb-TW-induced adaptations can be capitalized on for remediating spatial gait asymmetry, dynamic balance deficits, and impaired locomotor performance in adolescents with ULCP.