Background. This randomized controlled trial compared the diagnostic accuracy of the sonographic assessment of cervical length and clinical digital examination of the cervix in the second trimester regarding the prediction of preterm delivery in a low-risk population. Methods. In total, 282 unselected, asymptomatic women with singleton pregnancy randomly underwent sonographic cervical length measurement (study group, n = 138) or clinical digital examination (control group, n = 144) in the second trimester. In the study group cervical length ≤5th percentile (≤24 mm) for our population was defined as shortened. In the control group, Bishop score ≥95th percentile (≥4) for our population was defined as high. The primary outcome measure was the diagnostic accuracy of both tests regarding the prediction of preterm delivery (<37 weeks). Results. Shortened cervical length was found in 6/138 (4.3%) women whereas the high Bishop score was found in 17/144 (11.8%) (p = 0.038, Fisher's exact test). The incidence of preterm delivery was 5.7% (16/282). Regarding the prediction of preterm delivery, shortened cervical length and high Bishop score had sensitivity 57.1% versus 33.3% and positive predictive value 66.7% versus 17.6%. Shortened cervical length in comparison with high Bishop score had 12-fold higher positive likelihood ratio for preterm delivery in a low-risk population (37.4; 95%CI [8.2–170.7] versus 3.2; 95%CI [1.1–9.2]). Conclusion. Sonographic assessment of cervical length has better diagnostic accuracy in the prediction of preterm delivery compared to digital examination in a low-risk population.