Objective There is a dilemma of ibuprofen treatment with patent ductus arteriosus (PDA) as to how and when to treat. We aimed to clarify this issue in very preterm infants (VPIs; < 32 weeks). Methods: This retrospective study included 1 659 VPIs who were diagnosed with PDA according to echocardiographic examinations and cardiovascular dysfunction scoring system (the CVD scoring). The VPIs were classified into six groups (A1, A2, A3, B1, B2, and B3) based on CVD scores (A, < 3, and B, ≥ 3), and treatment with ibuprofen for PDA (1, conservational management; 2, early ibuprofen treatment; and 3, late ibuprofen treatment). Treatment was stopped when PDA was closed, CVD score was zero or PDA needed ligation. Results: VPIs with CVD scores < 3 had most PDA closure without surgery, and early ibuprofen treatment did not significantly affect PDA closure. VPIs with CVD scores ≥ 3 had some PDA closure after 2 courses of treatment, but closure rates decreased linearly with ibuprofen course (1^st 75.2%, 2^nd 62.3%, 3^rd 50.0%, P < 0.0001), and early ibuprofen treatment (group B2) did not increase PDA closure compared to late ibuprofen treatment (group B3). In these same infants, the longer they were in CVD scores ≥ 3, the more the complications of preterm were increased (retinopathy of prematurity ROP 1^st 16.5%, 2^nd 23.8%, 3^rd 29.6%, P = 0.016; bronchopulmonary dysplasia BPD 1^st 15.5%, 2^nd 26.7%, 3^rd 33.8%, P < 0.0001; intraventricular hemorrhage IVH 1^st 20.4%, 2^nd 32.4%, 3^rd 23.8%, P = 0.015). Conclusion: Ibuprofen is suggested for PDA closure when the PDA reopens or has developed into the stage when the CVD score ≥ 3.