A Hand Gesture (HG) constitutes a mode of nonverbal communication or non-vocal interaction, where discernible bodily movements convey specific messages, either independently or in combination with spoken language. Gesture recognition techniques are employed to create systems capable of transmitting information among individuals with disabilities for operating devices. In this paper, we propose the use of Electromyography (EMG) signals for classifying hand gestures through Employing both a Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) in combination is being used. The main problem lies in the noise signals produced by the EMG sensor. In this work, we employ a Low Pass Filter for preprocessing the signal images, followed by feature extraction using the CNN model. To reduce the high-dimensional signal and normalize the features for the classification process, we utilize LSTM. This experiment was conducted using the Ninapro DB1 dataset. This model successfully addresses the previous issues and attains an overall accuracy of 95.56%, Precision 91.90%, Recall 89.50%, and F1 score 85.50% 91.90%, then the accurate classification compares to other existing models like Bidirectional Convolutional Gated Recurrent Unit, and Deformable convolutional network models.