Aim This study evaluated whether the effect of submandibular gland botulinum neurotoxin A (BoNT‐A) injection can predict the outcome of submandibular duct relocation with sublingual gland excision (SMDR) in children with drooling. Furthermore, we compared the effectiveness of both procedures. Method A retrospective cohort study was performed in 42 children and adolescents (25 males, 17 females; mean [SD] age at BoNT‐A injection 11y [4], range 4–20y; mean [SD] age at SMDR 15y [4], range 7–23y) with cerebral palsy or another non‐progressive developmental disability who had undergone both BoNT‐A injection and SMDR for drooling. Main outcomes were the drooling quotient and the visual analogue scale (VAS) on drooling severity at 8 weeks and 32 weeks follow‐up. Results Failure or success of previous BoNT‐A injections had no influence on success of consecutive SMDR. Relative change in main outcomes showed no significant relation between BoNT‐A injection and SMDR for any follow‐up measurement. After 8 weeks, SMDR was more successful than BoNT‐A injection in diminishing VAS (VAS 80.0% vs 54.3%; drooling quotient 56.2% vs 51.0%). After 32 weeks, both drooling quotient (64.3% vs 29.5%) and VAS (75.7% vs 37.1%) showed significantly higher proportions of success for SMDR. Interpretation The effect of submandibular BoNT‐A injection does not predict subsequent SMDR success in drooling. Furthermore, SMDR has a larger and longer‐lasting positive effect on drooling than BoNT‐A injections. What this paper adds Submandibular botulinum neurotoxin A (BoNT‐A) injection effect does not predict submandibular duct relocation with sublingual gland excision outcome. Submandibular duct relocation is more effective and more permanent than BoNT‐A injection.