The aim of this study was to investigate whether early numeracy skills of South African first graders who are at-risk for mathematical learning difficulties can be improved with an intervention program. The participants were 267 children from 17 classrooms in the greater Johannesburg area. In this quasi-experimental small group intervention study (15 sessions over 5 weeks) the outcome measure was early numeracy skills. Based on pretest early numeracy scores, the children were divided into an intervention group (N = 40), a low performing control group (N = 32), and an average performing control group (N =195). The main result was that the intervention group had improved more in numerical relational skills, compared to low-controls; this effect remained statistically significant after controlling for executive functions, language skills and kindergarten attendance, and was also observable in the delayed post-measurement. Executive functions, language skills and kindergarten attendance all predicted the levelof early numeracy skills at the beginning of the intervention, but only executive functions explained individual differences in counting skills development from pre- to delayed posttest.