Objective: To compare the efficacy and safety of mycophenolate mofetil versus cyclosporine A in treating children with primary refractory nephrotic syndrome. Methods: Conducted a prospective randomized controlled clinical trial in 62 pediatric patients (including 44 boys and 18 girls), age ranged from 2.1 to 17.0 years; 32 cases presented with frequently relapsing nephrotic syndrome (FRNS) and 30 cases presented with steroid-resistant nephrotic syndrome (SRNS), who were admitted to department of Nephrology, Children's Hospital Affiliated to Capital Institute of Pediatrics from October 2013 to October 2015. The patients received either mycophenolate mofetil (20-30)mg/(kg·d) or cyclosporine A (3-5)mg/(kg·d) randomly, on the basis of prednisone treatment. Follow-up interview was conducted regularly for at least one year. Efficacy rate, relapse rate, time required for induction of remission, relapse-free period and prednisone dosage were compared between the two groups. Results: (1) Renal histologic examination, which was available for 17 patients, revealed minimal change disease in 8 patients, mesangial proliferative glomerulonephritis (MsPGN) in five, membranous nephropathy in two, and focal segmental glomerulosclerosis (FSGS) in two. (2) Comparison of mycophenolate mofetil versus cyclosporine A in children with FRNS: There were 14 patients with FRNS in mycophenolate mofetil group and 18 patients with FRNS in cyclosporine A group respectively. The relapse rate (episodes/year) in cyclosporine A group was lower than that of mycophenolate mofetil group (1.0 (0.0, 1.0) vs. 1.0 (1.0, 3.0), Z=-2.405, P=0.016). The relapse-free period (months) in cyclosporine A group was longer than that of mycophenolate mofetil group (10.0 (5.7, 12.1) vs. 5.0 (1.0, 11.0), Z=-1.984, P=0.047). No significant difference in dosage of prednisone was found between cyclosporine A and mycophenolate mofetil groups when followed up for 1 year. (3) Comparison of mycophenolate mofetil versus cyclosporine A in children with SRNS: The efficacy rate was 6/14 in mycophenolate mofetil group and 13/16 in cyclosporine A group. The complete remission rate was 4/14 in mycophenolate mofetil group and 12/16 in cyclosporine A group (P<0.05). The time (months) required for induction of remission in cyclosporine A group was significantly shorter than that of mycophenolate mofetil group (1.0 (1.0, 2.0) vs. 3.0 (2.5, 4.0), Z=-2.529, P=0.011). No significant differences were found between the two groups with respect to relapse-free period and relapse rate. (4) Except that one patient developed hypertensive encephalopathy in cyclosporine A group, no other serious adverse events were recorded. There were no significant differences between two groups with respect to adverse events. Conclusion: Our results indicated that both mycophenolate mofetil and cyclosporine A were effective in the treatment of children with refractory nephrotic syndrome. Cyclosporine A was superior to mycophenolate mofetil in preventing relapses in patients with FRNS and inducing complete remission in patients with SRNS. Although most patients were able to tolerate mycophenolate mofetil and cyclosporine A, but the toxicity and safety of cyclosporine A should be monitored closely.