Convolutional Neural Network (CNN) has targeted a significant evolution, specifically in the problem of identification of physical activities performed by human issues which are stated as Human Activity Recognition (HAR). In this study, a CNN deep learning model has been proposed in order to increase the recognition accuracy of different human activities. The innovations in CNN can be characterized in various ways including learning algorithms, optimization, activation function, regularization, and improvements in architecture. The following optimizers have been implemented with respect to the appropriate suggested layering structure for the proposed CNN model; SGD, Adam, Nadam, Adamax, AdaDelta, AdaGrad, and RMSProp. The proposed model uses raw data acquired from a set of inertial sensor and exploring numerous combinations of human activities; sitting, standing, jogging, both up and downstairs, and walking. In addition, evaluation process has targeted estimating the precision, f1-score, recall, accuracy and the confusion matrix. Experimental results show that the proposed CNN has achieved about 97.5% with respect to the Adam optimizer which would be considered as the most effectively recognizer compared to other deep learning architectures.