This paper presents a Deep Convolutional Neural Network (DCNN) model with a SoftMax classifier for the diagnosis of bearing and gear faults based on 2D grayscale images converted from raw vibration signals. The proposed model is validated and verified using a vibration dataset consisting of gear and bearing vibrations recorded under different operating conditions and fault conditions. The experimental results demonstrate that the DCNN achieves remarkable accuracies for both bearing fault diagnosis under different operating conditions and fault conditions, with high validation accuracy achieved. Furthermore, the performance of the proposed model is evaluated and compared with existing models in the literature for the same datasets, and it is concluded that the proposed DCNN is more efficient in terms of fault diagnosis accuracy, demonstrating its superiority. Therefore, the proposed high-accuracy DCNN can be an effective and accurate tool for fault diagnosis in various industrial applications.