In the real world detecting a facial emotion is challenging and complicated. To identify the subtle differences in facial expressions, Facial Emotion Recognition (FER) requires the model to learn. For image recognition tasks, a convolutional neural network (CNN) is a type of deep learning model that is commonly used. CNNs are able to learn features from images that are relevant to the task at hand, such as facial expressions. A pre-trained CNN is a CNN that has already been trained on a large dataset of images for another task, such as image classification. Pre-trained CNNs can be used to improve the performance of CNNs for other tasks, such as facial emotion recognition. The main difference between a CNN and a pre-trainedCNN is that a pre-trained CNN has already learned to extract features from images that are relevant to the task at hand. This means that a pre-trained CNN can be used to improve the performance of a CNN for the task at hand without having to train the CNN from scratch. Here we use MOBILENET as the pre-trained convolution neural network used with the help of the transfer learning technique. MOBILENET is a pre-trained CNN for FER, because it is efficient and accurate.EmoNet is a proposed mobile facial expression recognition system that utilizes the power of transfer learning and the efficiency of the MOBILENET model. The system aims to accurately classify facial expressions in real-time on mobile devices, making it accessible and user-friendly. The data is collected, pre-processed, and fed into the MOBILENET model for feature extraction. Stochastic gradient descent (SGD) is employed to train the pre-processed model, and its performance is evaluated using precision, recall, F1-measure, and accuracy metrics. Through experimental analysis and performance visualization, EmoNet demonstrates high estimation values and superior severity-level classification results compared to other models. This system offers a promising solution for efficient and accurate facial expression recognition, with potential applications in various domains, including emotion detection, human-computer interaction, and social robotics.