In real-time monitoring systems, strawberry freshness is a crucial metric for guaranteeing the quality of the product. This work investigates the application of transfer learning methods to effectively assess the freshness of strawberries, offering a less expensive option for creating deep neural networks (DNNs) from the ground up. We implemented multiple cutting-edge transfer learning models in a suggested real-time system architecture using a brand-new strawberry dataset. The Xception model outperformed the others in terms of accuracy, precision, recall, and F1-score, indicating its promise for a reliable and effective assessment of strawberry freshness. According to the research, transfer learning models provide a workable method for determining the freshness of fruit, saving time and money in real-time applications.