A prospective cohort study of a Chinese population was conducted to investigate the relationship between prenatal phthalates exposure and maternal hemoglobin or anemia. Based on the Ma'anshan Birth Cohort study, 7 phthalate metabolites were quantified in spot pregnancy urine samples (n = 9263) from 3269 pregnant women during each trimester. The maternal hemoglobin concentrations were obtained from electronic medical records at the same three time points for each participant during pregnancy. Anemia was defined as a hemoglobin concentration below 110 g/L in pregnant women. Repeated measures and trimester-specific analyses were used to estimate the effects of phthalates exposure on maternal hemoglobin and anemia. The prevalence of anemia was 3.6%, 27.0%, and 26.5% during the first, second and third trimester, respectively. Repeated measures analysis showed that hemoglobin concentrations decreased by 0.55, 0.19, 0.57, 0.49, and 0.54 g/L with each 1 ln-transformed concentration increase of MBP, MBzP, MEHP, MEOHP, and MEHHP, respectively. Exposure to MMP, MBP, MEHP, MEOHP, and MEHHP increased the risk of anemia by 1.11-fold, 1.21-fold, 1.20-fold, 1.13-fold, and 1.16-fold, respectively. Trimester-specific regression models stratified by the sample collection time during pregnancy generated consistent results. This is the first study focusing on the effect of prenatal phthalate exposures on hemoglobin or anemia in pregnant Chinese women. We found that prenatal phthalates exposure not only decreased the concentrations of hemoglobin but also showed associations with the prevalence of anemia. Associations appeared stronger for the subsample representing women pregnant with a male fetus than those pregnant with a female fetus. Anemia remains a moderate public health problem in China, and effective measures should be implemented.