Purpose This study aimed to examine the success rate of repeat endoscopic third ventriculostomy (redo-ETV) according to pattern of ventriculostoma closure based on observations in 97 paediatric redo-ETV patients. Methods Clinical data and intraoperative video recordings of 97 paediatric hydrocephalus patients who underwent redo-ETV due to ventriculostoma closure at two institutions were retrospectively analysed. We excluded patients with a history of intraventricular haemorrhage, cerebrospinal fluid (CSF) infection or CSF shunt surgery and those with incompletely penetrated membranes during the initial ETV. Results Verification of ventriculostoma closure was confirmed with cine phase-contrast magnetic resonance imaging and classified into 3 types: type 1, total closure of the ventriculostoma by gliosis or scar tissue that results in a non-translucent/opaque third ventricle floor; type 2, narrowing/closure of the ventriculostoma by newly formed translucent/semi-transparent membranes; and type 3, presence of a patent ventriculostoma orifice with CSF flow blockage by newly formed reactive membranes or arachnoidal webs in the basal cisterns. The overall success rate of redo-ETV was 37.1%. The success rates of redo-ETV according to closure type were 25% for type 1, 43.6% for type 2 and 38.2% for type 3. The frequency of type 1 ventriculostoma closure was significantly higher in patients with myelomeningocele-related hydrocephalus. Conclusion For patients with ventriculostoma closure after ETV, reopening of the stoma can be performed. Our findings regarding the frequencies of ventriculostoma closure types and the success rate of redo-ETV in paediatric patients according to ventriculostoma closure type are preliminary and should be verified by future studies.