ECG signals acquired from mobile devices by unskilled users are corrupted with several noises. Poor signal quality may result in an increased number of false alarms, degrading diagnostic performance, and increasing the burden on the doctors in decoding the information for further clinical intervention. So, it is necessary to assess the quality of the signals before doing any further processing. This paper presents a method for accessing the reliability of ECG signals obtained from wearable sensors. A morphological event-based quality assessment method is proposed where a signal will be classified as GOOD/BAD. Results show that our method can achieve an accuracy = 92 % with sensitivity = 0.98 and specificity = 0.59.