The non-invasive heart rate of the fetus is of great relevance for clinical measures to detect health fluctuations of the fetus during pregnancy. The electrocardiogram (ECG) is often used to detect cardiovascular malformations and heart disorders in the patient. An ECG is not able to measure fetal pulse rate because it shows a low signal-to-noise ratio due to the movement of the fetus in relation to midriff of the mother. The proposed work is based on effective removal of noise components, detection of normal sinus rhythms and arrhythmias of the fetus. Fluctuations in heart rate indicate the disease arrhythmia, i.e. either low or high heart rate. The main goal of this research is to provide an independent component analysis method for removing the maternal ECG(m-ECG) from the selected ECG signal to obtain the fetal ECG(f-ECG). Joint Approximation Diagonal Eigen Vector (JADE) is an algorithm used to discriminate the different components within the signal. The separated f-ECG is used for peak detection and feature extraction. The arrhythmia is detected from the extracted features. The proposed technique has been evaluated with 17 samples and an accuracy of 94.11% is achieved. © 2020 IEEE.