Vitamin A deficiency (VAD) and anemia are the most prevalent nutritional deficiency in children globally. The dried blood spot (DBS) method has been used in prevalence studies of VAD and anemia in different age groups. However, it has not yet been validated for children. This study aimed to assess the reproducibility and validity of DBS in the diagnosis of VAD and anemia in preschoolers. Venous and capillary blood samples were collected from a representative sample of children <5 y old who attended the public health system in Rio de Janeiro. Serum retinol and hemoglobin were measured in 235 and 182 children, respectively. Serum retinol was measured with HPLC and hemoglobin was measured with spectrophotometry in samples of venous (gold standard) and capillary blood (test method, DBS). DBS reproducibility was assessed with the intraclass correlation coefficient (ICC), κ, and prevalence-adjusted and bias-adjusted κ (PABAK). DBS validity was assessed with sensitivity, specificity, accuracy index (AI), positive predictive value (PPV), and negative predictive value (NPV). DBS showed very good reproducibility for serum retinol (ICC = 0.94, κ = 0.83, PABAK = 0.76) and very good/good reproducibility for hemoglobin (ICC = 0.86, κ = 0.69, PABAK = 0.69). Prevalence rates for VAD by the reference and test methods were 11.5% and 11.9%, respectively, whereas the anemia rates were 19.2% and 46.2%. The test method showed low sensitivity (33%) and PPV (32%) and high specificity (91%) and NPV (92%) for serum retinol. For hemoglobin, the test method showed fair sensitivity (71%), low PPV (30%), fair specificity (60%), and high NPV (90%). AI was 83% for VAD and 62% for anemia. The results suggest that DBS is adequate for the diagnosis of VAD in preschool children, but not for anemia.