Monkeypox is a viral infectious disease that presents a significant public health concern, particularly in West and Central Africa. The timely detection and precise diagnosis of Monkeypox are vital for preventing disease transmission and controlling outbreaks. However, in disease-dense areas, diagnostic methods are limited, resulting in delayed detection and contributing to the disease's spread. This study focuses on conducting a comparative analysis of leveraging deep learning Convolutional Neural Network (CNN) models to detect and predict Monkeypox based on skin lesion images. This study also addresses the challenge of overfitting commonly encountered in CNN models due to dense layers and restricted image datasets. Various techniques are applied to minimize overfitting while maintaining a high accuracy percentage. Additionally, the performance of the Monkeypox model, InceptionV3, and MobileNetV2 CNN models are compared to determine the most effective approach for Monkeypox identification. The evaluation results suggest that the MobileNetV2 technique performs best without overfitting with an accuracy of 91% while detecting the Monkeypox virus. However, enhancing the detection capabilities of robust models necessitates an increased volume of training data to effectively train deep learning architectures.