Background Studies have shown a mixed association between socioeconomic status [SES] and prevalent HIV infection across and within settings in sub-Saharan Africa. In general, the relationship between years of formal education and HIV infection changed from a positive to a negative association with maturity of the HIV epidemic. Our objective was to determine the association between SES and HIV in women of reproductive age in the Free State [FSP] and Western Cape Provinces [WCP] of South Africa [SA]. Study design Cross sectional Setting South Africa Methods We conducted secondary analysis on 1906 women of reproductive age from a 2007-2008 survey that evaluated effectiveness of Prevention of Mother-to-Child HIV Transmission programs. SES was measured by household wealth quintiles, years of formal education and employment status. Our analysis principally utilized logistic regression for survey data. Results There was a significant negative trend between prevalent HIV infection and wealth quintile in WCP [p-value<0.001] and FSP [p-value=0.025]. In adjusted analysis, every additional year of formal education was associated with a 10% [adjusted Odds Ratio [aOR] 0.90 [95% Confidence Interval [CI]: 0.85-0.96] significant reduction in risk of prevalent HIV infection in WCP but no significant association was observed in FSP [aOR 0.99 CI: 0.89-1.11]. There was no significant association between employment and prevalent HIV in each province: [aOR 1.54 CI: 0.84-2.84] in WCP and [aOR 0.96 CI: 0.71-1.30] in FSP. Conclusion The association between HIV infection and SES differed by province and by measure of SES and underscores the disproportionately higher burden of prevalent HIV infection among poorer and lowly educated women. Our findings suggest the need for re-evaluation of whether current HIV prevention efforts meet needs of the least educated [in WCP] and the poorest women [both WCP and FSP], and point to the need to investigate additional or tailored strategies for these women.