HIV viral load is more reliable tool for monitoring treatment throughout the course of HIV/AIDS, but the test may be expensive in resource-limited settings. Therefore, enumeration of CD4 T-lymphocyte count remains important in these settings. This study evaluated the performance of BDFACSPresto, a near-patient CD4 counter planned to be used in primary healthcare clinics in Thailand. Results of percent, absolute CD4 count and hemoglobin (Hb) on the FACSPresto were compared with the TriTEST/TruCOUNT/BDFACSCalibur method and a Sysmex hematology analyzer. Phase I of the study was performed in an ISO15189 laboratory. Both percentage and absolute values showed Passing–Bablok slopes within 0.98–1.06 and 0.97–1.13, mean Bland–Altman biases of +1.2% and +20.5 cells/µL, respectively. In phase II, venous and some capillary blood samples were analyzed in four primary healthcare clinics. The results showed good correlation between capillary and venous blood. For venous blood samples, regression lines showed slopes of 1.01–1.05 and 1.01–1.07 for all percentage and absolute values. The overall mean biases were +0.9% and +17.0 cells/µL. For Hb, Passing–Bablok regression result gave slope within 1.01–1.07 and mean bias of −0.06 g/dL. Thus, CD4 enumeration in blood by the FACSPresto is reliable and can be performed to an identical standard at primary healthcare clinics.