To clarify whether twin fetuses could be clustered quantitatively observing in utero cardiovascular system function, 71 twin-paired fetuses between 24 and 40 weeks' gestation were subjected. Cluster analysis was performed using four variables; end-diastolic dimension and fractional shortening in the right ventricle (EDD, FS), resistance index in the umbilical artery (RIUA) and urine production rate (UPR). Three different groups were revealed. In Cluster 1 (C1), larger fetuses showed a significant increase in both EDD and UPR, and a significant decrease in FS, while smaller fetuses had a significant increase in RIUA and a decrease in UPR. In Cluster 2 (C2), smaller fetuses showed a significant increase in RIUA. Cluster 3 (C3) showed no significant deviations in any variable when referring to singleton pregnancies. C1 had a higher perinatal mortality rate than either C2 or C3, with the cause of death being cardiac failure in larger fetuses and renal failure in smaller fetuses. The frequencies of discordancy, hemoglobin difference above 5 g/dl and monochorionic placenta in C1 were also higher than fetuses in C3. C2 also showed higher frequencies of discordancy and small-for-gestational age than C3 fetuses. In C2 and C3, respiratory insufficiency was found to be the cause of neonatal death. The results revealed that C1 consists of twins with both larger and smaller fetuses displaying cardiac failure with an increase in renal perfusion and a decrease in renal and placental perfusion, respectively, and C2 consists of twins with smaller fetuses having both placental insufficiency and intrauterine growth retardation. C3 consists of twins with both lager and smaller fetuses having normal circulation, when compared to singleton pregnancy.