To evaluate retrospectively the prevalence of gestational diabetes (GD) in pregnancies obtained with myo-inositol administration in women with polycystic ovary syndrome. A total of 98 pregnancies in PCOS women obtained in a 3-year period, either with myo-inositol (n. 54), or with metformin (n. 44) were considered. While myo-inositol was assumed through the whole pregnancy, the group of women treated with metformin stopped the drug assumption after pregnancy diagnosis, and was considered as a control group. After having eliminated cases of miscarriages and twin pregnancies, a definitive number of 46 women in the myo-inositol group and 37 in the control group was taken in account to be retrospectively evaluated. The primary outcome measure was GD occurrence in both groups; whereas secondary outcome measures were pregnancy outcomes: hypertensive disorders, pre-term birth, macrosomia and caesarean section occurrence. Prevalence of GD in the myo-inositol group was 17.4% versus 54% in the control group, with a highly significant difference also after adjusting for covariates. Consequently, in the control group the risk of GD occurrence was more than double compared to the myo-inositol group, with an odds ratio 2.4 (confidence interval 95%, 1.3–4.4). There was no difference between the groups in relation to secondary outcome measures. This study suggests a possible effect of myo-inositol in the primary prevention of GD in PCOS women.