Background: A three-arm parallel group, randomised controlled trial set in general dental practices in England, Scotland, and Wales was undertaken to evaluate three strategies to manage dental caries in primary teeth. Children, with at least one primary molar with caries into dentine, were randomised to receive Conventional with best practice prevention (C + P), Biological with best practice prevention (B + P), or best practice Prevention Alone (PA). Methods: Data on costs were collected via case report forms completed by clinical staff at every visit. The coprimary outcomes were incidence of, and number of episodes of, dental pain and/or infection avoided. The three strategies were ranked in order of mean cost and a more costly strategy was compared with a less costly strategy in terms of incremental cost-effectiveness. Costs and outcomes were discounted at 3.5%. Results: A total of 1144 children were randomised with data on 1058 children (C + P n = 352, B + P n = 352, PA n = 354) used in the analysis. On average, it costs £230 to manage dental caries in primary teeth over a period of up to 36 months. Managing children in PA was, on average, £19 (97.5% CI:-£18 to £55) less costly than managing those in B + P. In terms of effectiveness, on average, there were fewer incidences of, (− 0.06; 97.5% CI: − 0.14 to 0.02) and fewer episodes of dental pain and/or infection (− 0.14; 97.5% CI: − 0.29 to 0.71) in B + P compared to PA. C + P was unlikely to be considered cost-effective, as it was more costly and less effective than B + P. Conclusions: The mean cost of a child avoiding any dental pain and/or infection (incidence) was £330 and the mean cost per episode of dental pain and/or infection avoided was £130. At these thresholds B + P has the highest probability of being considered cost-effective. Over the willingness to pay thresholds considered, the probability of B + P being considered cost-effective never exceeded 75%. Trial registration: The trial was prospectively registered with the ISRCTN (reference number ISRCTN77044005) on the 26th January 2009 and East of Scotland Research Ethics Committee provided ethical approved (REC reference: 12/ES/0047).