Electromyography signals usage has grown in the scientific community and in treatment and disease diagnoses, such as prosthesis and orthosis control. A lot of these systems use gesture classification to take and execute actions. The first step of the caption of the surface electromyography signal (sEMG) uses an electrode for the caption of the signal, then this signal is amplified, and in its last stage, its processing. The most common electrode used is the Ag/AgCl electrode. However, not all users can use it due to their skin sensitivity or allergies. In front of this problem, this case study looks for the viability of changing AgCl electrodes with dry electrodes in a system where a gesture classification based on sEMG signals is used to execute actions. To validate this hypothesis, the sEMG signal was taken from 14 volunteers with two types of electrodes: Ag/AgCl electrode and a dry electrode built with the Medtex P180 mesh. An SVM classifier was built to classify three gestures (wrist extension, wrist flexion, and palm grip). Six combinations to train the model were explored to ensure the possibility of electrode replacement, and their metrics were obtained as the average of the Leave-One-Subject-Out (LOSO) methodology. The analysis of the power ratio relation between the signals from these two types of electrodes was also used to understate the results and quality of the sEMG signal. The best model obtained (accuracy of 95.47%) was trained and tested with AgCl sEMG data from Ag/AgCl electrodes, and the worst model was trained and tested with data from dry electrodes (accuracy of 83.45%). Using Friedman’s hypotheses and other analyses, like the confusion matrix, it was possible to confirm the viability of replacing one electrode with the other by assuming the possibility of a performance reduction of the model.