Prematurity is one of major health concerns that can cause short-term and long-term complications. There are evidences that uterine electromyography (EMG), also called electrohysterography (EHG), measuring the electrical activity of uterine muscles is a valuable diagnosis for term and preterm labor assessments. In this study, seventeen time-domain features commonly used in EMG signal processing and analysis of EHG data are examined and applied for preterm-term birth classifications. Two feature selection methods including Pearson’s correlation coefficient and p-value of t-test are also applied to reduce the dimension of features. The classifications are performed using support vector machine (SVM) classifiers with polynomial kernel function. From the computational results, the best performance on pretermterm birth classifications of selected time-domain features of EHG data determined from the product of sensitivity and specificity achieved is the accuracy of 0.6667, the sensitivity of 0.7895 and the specificity of 0.6503. In addition, the computational results suggest that the remarkable time-domain features of EHG data for the preterm-term classifications are the difference absolute standard deviation value, the waveform length, the average amplitude change and the ν-order.