As a severe morbidity during pregnancy, the etiology of spontaneous pregnancy loss (SPL) remains largely unknown. Serum glycated hemoglobin (HbA1c) level is an established predictor of SPL risk among women with diabetes, but little is known about whether such an association exists among pregnant women without diabetes when glycemic levels are within the normal range. This study aimed to quantify the association between maternal HbA1c levels in early pregnancy and subsequent SPL risk in a cohort of pregnant women without diabetes. This prospective cohort study involved 10,773 pregnant women without diabetes enrolled at their first antenatal care visit at a hospital's early pregnancy clinic from March 2016 to December 2018 in Shanghai, China. HbA1c and fasting blood glucose (FBG) levels were examined at enrollment. Participants with diabetes before or pregnancy or those diagnosed with gestational diabetes were excluded. Diagnosis of SPL, defined as fetal death occurring before 28 gestational weeks, was derived from medical records and confirmed via telephone interviews. We used generalized linear models to quantify the associations of continuous and dichotomized maternal HbA1c levels with SPL risk and reported crude and adjusted risk ratios (RRs) and 95% CIs. A restricted cubic spline (RCS) regression model was used to assess the potential nonlinear dose-response relationship. Adjusted covariates included maternal age, education level, preconception BMI, gestational weeks, gravidity, history of adverse pregnancy outcomes, family history of diabetes, folic acid supplementation, and smoking and drinking during the periconception period. In total, 273 (2.5%) SPL cases occurred. Every 0.5% increase in HbA1c levels was linearly associated with a 23% increase in SPL risk (adjusted RR [aRR] 1.23; 95% CI 1.01-1.50). The RCS model revealed that this association was linear (P=.77 for the nonlinearity test). Analyses based on dichotomized HbA1c levels showed a significantly increased risk of SPL when HbA1c levels were ≥5.9% (aRR 1.67; 95% CI 0.67-3.67), and the significance threshold was ≥5.6% (aRR 1.60; 95% CI 1.01-2.54). Sensitivity analyses showed similar results when including the participants with missing SPL records or HbA1c data. Linear associations of HbA1c levels remained significant even in the subgroups without overweight, alcohol consumption, and a family history of diabetes and adverse pregnancy outcomes. Every 1 mmol/L increment in maternal FBG levels was associated with a >2-fold higher risk of SPL (aRR 2.12; 95% CI 1.61-2.80; P<.001). Higher HbA1c levels in early pregnant women without diabetes are associated with an increased SPL risk in a dose-response manner. Pregnant women with an HbA1c level above 5.6% at early gestation need attention for its potentially increased risk for SPL. Our findings support the need to monitor HbA1c levels to identify individuals at high risk of subsequent SPL in the general population of pregnant women. ClinicalTrials.gov NCT02737644; https://clinicaltrials.gov/study/NCT02737644.