The date palm, which occupies 50% of the nation's total agricultural land and 80% of all fruit crops, is Oman's most significant agricultural crop. Technology advancement has benefitted date palm pruning and pollination for increased yield and harvesting. Like that, introducing technology for date fruit post-harvest processing to Oman could have a significant economic impact. Widely used manual date fruit classification based on quality is tedious, inconsistent, and time-consuming. Another issue with manual processing is the non-uniformity of sorting since human sensors' quality and calibration vary from person to person. This paper proposes a transfer-learning-based automatic date fruit classification system. The model included an electromechanical system for the steady movement of the date fruits. A microcontroller interfaced with a sensor identifies the presence of date fruits on the conveyer belt and initiates control signals for the fruit image capturing. The captured images are categorized using the transfer learning-based models incorporated with the MIT App and the teachable machine. A microcontroller-based system receives commands from the transfer learning-based system and activates piston movement. Piston movement is employed to collect categorized dates in smart bins. The proposed model could achieve appreciable classification accuracy.