Background: Hemoglobulinopathies (Thalassemia, Sickle cell anemia) are an important public health challenge worldwide with an estimated number of 330,000 affected newborns annually. Delay in diagnosis not only increases the morbidity and mortality rates, but also can lead to uncertainties about success of thalassemia prevention program (TPP). This study is the first to determine the delay in thalassemia diagnosis in Iranian population. Methods: This registry-based cross-sectional study was conducted on 1003 enrolled thalassemic patients in 2015 in Shiraz, Iran. Univariate and multivariate ordinal logistic regression models were used to assess the factors associated with delay in diagnosis. Results: Of 981 patients, 48.5% were female, 71.2% were thalassemia major, and 23.8% were death cases. The delay in diagnosis was observed among 64.9% of the patients with a mean of 13.4 months (95% CI: 10.9, 15.9). Multivariate ordinal logistic regression showed that girls (adjusted OR = 1.32), and dead patients (adjusted OR = 1.95) were more likely to have delayed diagnosis. There was an increasing trend of risk in delayed diagnosis associated with one-year per birth cohort. The ORs were 1.0, 1.52, 1.55, and 2.22, for birth cohort 1980 and earlier, 1981 -1990, 1991 -2000, and 2001 to the present, respectively (P for trend = 0.014). In addition, the odds of delayed diagnosis in thalassemia major patients were significantly 0.58 times lower than those with thalassemia intermedia. Conclusions: A high proportion of delayed diagnosis was found. These results could explain the poor outcomes for thalassemia patients. Educational programs for community and revising the TPP are required for early detection of the disease.