Background <p>Effective sedation for paediatric magnetic resonance imaging (MRI) requires balancing depth control and safety. This dose-finding study aims to establish the 50% (ED₅₀) and 90% (ED₉₀) effective doses of intravenous remimazolam following intranasal dexmedetomidine premedication to optimise this sequential regimen.</p> Methods <p>In this single-centre trial (March-December 2024), children aged 3 months to 6 years requiring MRI sedation were enrolled. After intranasal dexmedetomidine (3&#xa0;µg·kg⁻¹), children adequately sedated by dexmedetomidine alone were excluded. 150 children with inadequate sedation were stratified by age (3–12 months, 1–3 years, 3–6 years) and administered remimazolam via Dixon’s up-and-down method. ED₅₀ and ED₉₀ with 95%CI were calculated via probit regression. Vital signs, adverse events, awake time, and post-sedation side effects were recorded.</p> Results <p>Remimazolam for successful sedation maintained consistent ED₅₀ values of 0.12&#xa0;mg·kg^− 1 across all paediatric age groups. ED₉₀ values slightly increased from 0.20&#xa0;mg·kg^− 1 in infants and toddlers to 0.23&#xa0;mg·kg^− 1 in preschoolers. MRI was successfully completed in 93.3% of patients, while 6.7% required additional medication during the examination. Adverse events occurred in 15.3% of children, including transient hypotension in 2%, bradycardia in 0.7%, and hypoxia in 0.7%; all events were self-limiting without intervention.</p> Conclusions <p>This prospective study defines the ED₅₀ and ED₉₀ of remimazolam following dexmedetomidine premedication in dexmedetomidine-non-responders for paediatric MRI sedation. The sequential regimen demonstrated consistent weight-based dosing requirements with a favorable safety profile.</p> Trial registration <p>The trial was prospectively registered in the Chinese Clinical Trial Registry (chictr.org.cn; No. ChiCTR2400081218) on February 26, 2024.</p>