Electromyography (EMG) is a technique used to assess and record the electrical activity produced by skeletal muscles. This information can be used to diagnose muscle disorders, such as myopathy and Amyotrophic Lateral Sclerosis (ALS). In this study, we made a significant contribution to the field by proposing an automated method for classifying EMG signals that is more accurate than previous methods. Our method uses tunable-Q factor wavelet transform (TQWT) to decompose the EMG signal into its constituent components. These components are then used to calculate seven features that characterize the signal which are Interquartile Range (IQR), Mean Absolute Value (MAV), Mode, Kurtosis, Standard Deviation, Ratio of the absolute mean value , and Skewness. The features are then used to train a Bagging ensemble classifier. We evaluated our method on a dataset of EMG signals from healthy people, patients with myopathy, and patients with ALS. Our method achieved an accuracy of 99% in classifying the EMG signals. Our results suggest that the proposed method is a promising approach for diagnosing muscle disorders using EMG. This method could be used to improve the early diagnosis and treatment of these disorders.