Deepfake images have emerged as a notable concern in today’s digital era. They pose significant threat to digital privacy, security and integrity. Deepfake images have profound impact on society. Deepfake images are used to create false content, which is then disseminated through social media apps. It is essential to detect them to combat the growing threat of deepfake technology. Deepfake image detection is a challenging issue, as advancements in artificial intelligence have led to the creation of highly realistic images that are difficult to identify. Several machine learning and convolutional neural network-based models have been proposed for deepfake image detection; however, their accuracy remains limited. In this paper, we propose a deepfake detection model based on MobileNetV2 to improve classification accuracy and efficiency. The proposed model is lightweight and efficient, making it well-suited for deepfake detection by effectively capturing complex patterns and inconsistencies in images. Our results and comparative analysis demonstrate that the proposed MobileNetV2-based model exhibits better performance than theexisting VGG16 and ResNet models in terms of accuracy.This work highlights the potential of MobileNetV2 in addressing the growing challenge of deepfake image detection.