Preliminary data suggest that strict glycemic control in twin pregnancies with gestational diabetes mellitus may not improve outcomes but might increase the risk of fetal growth restriction. This study aimed to investigate the association of maternal glycemic control with the risk of gestational diabetes mellitus-related complications and small for gestational age in twin pregnancies complicated by gestational diabetes mellitus. This was a retrospective cohort study of all patients with a twin pregnancy complicated by gestational diabetes mellitus in a single tertiary center between 2011 and 2020, and a matched control group of patients with a twin pregnancy without gestational diabetes mellitus in a 1:3 ratio. The exposure was the level of glycemic control, described as the proportion of fasting, postprandial, and overall glucose values within target. Good glycemic control was defined as a proportion of values within target above the 50th percentile. The first coprimary outcome was a composite variable of neonatal morbidity, defined as at least 1 of the following: birthweight >90th centile for gestational age, hypoglycemia requiring treatment, jaundice requiring phototherapy, birth trauma, or admission to the neonatal intensive care unit at term. A second coprimary outcome was small for gestational age, defined as birthweight <10th centile or <3rd centile for gestational age. Associations between the level of glycemic control and the study outcomes were estimated using logistic regression analysis and were expressed as adjusted odds ratio with 95% confidence interval. A total of 105 patients with gestational diabetes mellitus in a twin pregnancy met the study criteria. The overall rate of the primary outcome was 32.4% (34/105), and the overall proportion of pregnancies with a small for gestational age newborn at birth was 43.8% (46/105). Good glycemic control was not associated with a reduction in the risk of composite neonatal morbidity when compared with suboptimal glycemic control (32.1% vs 32.7%; adjusted odds ratio, 2.06 [95% confidence interval, 0.77-5.49]). However, good glycemic control was associated with higher odds of small for gestational age compared with nongestational diabetes mellitus pregnancies, especially in the subgroup of diet-treated gestational diabetes mellitus (65.5% vs 34.0%, respectively; adjusted odds ratio, 4.17 [95% confidence interval, 1.74-10.01] for small for gestational age <10th centile; and 24.1% vs 7.0%, respectively; adjusted odds ratio, 3.97 [95% confidence interval, 1.42-11.10] for small for gestational age <3rd centile). In contrast, the rate of small for gestational age in gestational diabetes mellitus pregnancies with suboptimal control was not considerably different when compared with non-gestational diabetes mellitus pregnancies. In addition, in cases of diet-treated gestational diabetes mellitus, good glycemic control was associated with a left-shift of the distribution of birthweight centiles, whereas the distribution of birthweight centiles among gestational diabetes mellitus pregnancies with suboptimal control was similar to that of nongestational diabetes mellitus pregnancies. In patients with gestational diabetes mellitus in a twin pregnancy, good glycemic control is not associated with a reduction in the risk of gestational diabetes mellitus-related complications but may increase the risk of a small for gestational age newborn in the subgroup of patients with mild (diet-treated) gestational diabetes mellitus. These findings further question whether the gestational diabetes mellitus glycemic targets used in singleton pregnancies also apply to twin pregnancies and support the concern that applying the same diagnostic criteria and glycemic targets in twin pregnancies may result in overdiagnosis and overtreatment of gestational diabetes mellitus and potential neonatal harm.