Objective: Clinical typing methods of the oral pathogen Streptococcus mutans with molecular analysis can be very specific, but expensive. Repetitive extragenic palindromic PCR (rep-PCR) is a relatively inexpensive pre-screening alternative for isolate selection for additional analyses. This study evaluated the prediction accuracy of using rep-PCR to identify S. mutans multilocus sequence typing (MLST) sequence types (ST) among children and their family members. Potential S. mutans strain sources were evaluated for evidence of transmission. Material and methods: Ten dendrograms (rep-PCR), with 20 isolates each of the 10 most common S. mutans genotypes, were generated from different subjects. Using a cut-off of 98% similarity, 7–11 isolates of each genotype were selected for MLST analysis to determine ST match/no-match. Results: Overall, rep-PCR was 75% effective at determining MLST ST match/no-match and 90% effective when applied to related individuals. Most genotypes were further differentiated by MLST. MLST ST diversity was greatest for one genotype (genotype 12, G12) and evidence of transmission among children and their family members was identified by rep-PCR and MLST. Younger children (6 months to 4 years old) shared ST with their mothers but 50% of older children (5–9 years old) had ST not identified in their mother. Six ST were shared between different families and probable source members were identified. Conclusion: This study confirms that rep-PCR offers an affordable option to predict diverse isolates for downstream applications. Clinical relevance: Using a combined rep-PCR and MLST approach, it is possible to track probable transmission and strain sources for S. mutans genotypes.