OBJECTIVE To compare two methods for Streptococcus mutans detection and quantification in the human oral cavity: a chairside commercial test and a molecular-based real-time quantitative polymerase chain reaction (qPCR) method. METHODS A total of 688 whole saliva samples were collected from 344 children aged 3 and 5 and their biological mothers. Caries status was examined using a World Health Organisation survey method. S. mutans levels were measured using the Dentocult SM Strip mutans test and scored as colony forming units per millilitre of saliva. Meanwhile, bacterial genomic DNA was extracted from the saliva, qPCR was performed with S. mutans species-specific primers, and absolute S. mutans DNA concentrations were obtained and scored as micrograms of DNA per millilitre of saliva. The two methods were compared for sensitivity, specificity, agreement and correlation with caries status. RESULTS Significantly more participants tested positive for S. mutans by qPCR than in the chairside SM Strip test (82.4% vs 71.4%). When only the highest and lowest test scores were considered, the agreement between the two methods assessing S. mutans colonisation was 0.956. Children with high levels of S. mutans in their saliva were six to eight times more likely to develop dental caries at 5 years old. CONCLUSIONS The study provides new evidence supporting the use of the chairside SM Strip test or the qPCR assay for the detection and quantification of S. mutans colonisation in saliva as the analytical approach of choice for caries risk assessment in clinical and epidemiological studies.