The objectives of the present study were to evaluate and compare the accuracy of transabdominal ultrasonography and pregnancy-associated glycoprotein (PAG) assay in the diagnosis of twin pregnancies in ewes and to evaluate the utility of the PAG assay for predicting fetal gender in singleton pregnancies. The animals in the study consisted of 179 pregnant ewes. The number of fetuses in the ewe was determined using transabdominal ultrasonography between days 40 and 60 (on days 40, 45, 50, 55, and 60). Blood samples were collected from all the ewes on the same day to determine the PAG concentrations. The results found were highly sensitive for the detection of twin pregnancies by transabdominal ultrasonography. The accuracy of transabdominal ultrasonography in detecting twin pregnancies was found to be higher on day 60 than on other days (P < 0.05). The sensitivities of PAG assay in detecting twin pregnancies on days 40, 45, 50, 55, and 60 were 91.67%, 66.67%, 81.82%, 88.89%, and 33.33%, respectively. The accuracies of the PAG assay in detecting twin pregnancies on days 40, 45, and 50 were found to be statistically significant higher than other days (P < 0.05). The PAG assay had low sensitivity and specificity for predicting fetal gender. It was concluded that twin pregnancies in ewes can be diagnosed with high accuracy using transabdominal ultrasonography on day 60 of pregnancy and as well as using the PAG assay during the early days of pregnancy (on days 40, 45, and 50). The PAG assay is not useful for predicting fetal gender.