As the Technology is raising and awareness is step forward in all domains and especially in medical science raised people precaution in all cases. Still in Gynecology we are facing issues with fatal birth cases. The primary issues is the stress bared by the fetal at the time of Labour. Cardiotocographic (CTG) monitor is a device which is used to recursively monitor the status of both the maternal pulse rate and fetal heart rate at the time of labour. In general due to lack of oxygen for the fetus which may be raised due to maternal uterus contraction. This may lead to abnormal issues and sometime raises to the risky case of the labour or sometimes lead to death as well for both the lives. Hence, by using classification techniques the abnormality features are diagnosed with the help of Cardiotocography (CTG) signal. The former classification methods are difficult to process the non-stationary data from CTG and the dataset imbalance. Here we are introducing a novel methodology for Time-frequency (TF) features from the raw data of CTG. This study makes use of the following to evaluate fetal distress: 1) DT 2) SVM 3) RF 4) NN 5) GB We present in this paper a new algorithm which improves the accuracy to 95.7% which is higher than what was obtained in previous research.