Variation in body weight (BW) of piglets at weaning is a drawback for successful implementation of multi-suckling (MS) systems. The current study investigated the combination of two intervention strategies, i.e. the timing of grouping and split-weaning, aiming to improve the BW gain of low birthweight piglets in an MS system and thereby reduce the BW variation at weaning on day 48 postpartum (p.p.). Eight batches of 5 sows with their litters were divided into 4 control (CTRL) and 4 treatment (TREAT) batches. In each litter, the second lowest (LBW) and highest (HBW) birthweight piglets from both sexes were selected as focal piglets. CTRL piglets were grouped on day 8 p.p. and no split-weaning was applied; TREAT piglets were grouped on day 13 p.p. and the three heaviest non-focal piglets per litter were split-weaned on day 35 p.p. Behaviour and feed intake were measured in focal piglets, and BW and mortality were measured in all piglets. Results showed that: (1) Throughout lactation there were no differences in BW or BW gain between CTRL and TREAT, nor were birthweight × treatment interactions found. (2) After grouping, there were no obvious differences between CTRL and TREAT in feeding and suckling behaviours on day 18, damage scores on snout, ear or tail and skin lesions on day 27, nor were birthweight × treatment interactions found. (3) After split-weaning, in week 6, piglets in TREAT tended to consume less feed than CTRL (P = 0.072). Low birthweight piglets in TREAT consumed numerically less feed and spent numerically less time contacting feed during the day than CTRL piglets. In week 6, there was a significant birthweight × treatment interaction in dry matter milk intake (P = 0.030), caused by a higher milk intake of TREAT-LBW piglets compared with CTRL-LBW piglets. In week 6, TREAT piglets tended to be present more at front and middle teats (P = 0.052) and tended to have lower snout damage scores (P = 0.084) than CTRL piglets. (4) Piglet crushing of all piglets in TREAT tended to be higher during the period when TREAT piglets were not yet grouped i.e. during days 9–14 than CTRL (P = 0.087). To conclude, split-weaning of the heavy piglets increased milk intake particularly of low birthweight piglets but this did not lead to a reduction in BW variation at weaning, as the increased milk intake was largely compensated for by a simultaneous decrease in feed intake.