Background: HIV-infected infants have high risk of death in the first two years of life if untreated. WHO guidelines recommend early infant HIV diagnosis (EID) of all HIV-exposed infants and immediate antiretroviral therapy (ART) in HIV-infected children under 24-months. We assessed the cost-effectiveness of this strategy in HIV-exposed non-breastfed children in Thailand. Methods: A decision analytic model of HIV diagnosis and disease progression compared: EID using DNA PCR with immediate ART (Early-Early); or EID with deferred ART based on immune/clinical criteria (Early-Late); vs. clinical/serology based diagnosis and deferred ART (Reference). The model was populated with survival and cost data from a Thai observational cohort and the literature. Incremental cost-effectiveness ratio per life-year gained (LYG) was compared against the Reference strategy. Costs and outcomes were discounted at 3%. Results: Mean discounted life expectancy of HIV-infected children increased from 13.3 years in the Reference strategy to 14.3 in the Early-Late and 17.8 years in Early-Early strategies. The mean discounted lifetime cost was $17,335, $22,583 and $29,108, respectively. The cost-effectiveness ratio of Early-Late and Early-Early strategies was $5,149 and $2,615 per LYG, respectively as compared to the Reference strategy. The Early-Early strategy was most cost-effective at approximately half the domestic product per capita per LYG ($ 4,420 in Thailand 2011). The results were robust in deterministic and probabilistic sensitivity analyses including varying perinatal transmission rates. Conclusion: In Thailand, EID and immediate ART would lead to major survival benefits and is cost- effective. These findings strongly support the adoption of WHO recommendations as routine care. © 2014 Collins et al.