Aiming at the problems that ham, sausage and other meat products are prone to quality spoilage and shortened shelf life in the process of production, transportation and sales, the quality spoilage parameters of ham were studied based on nearinfrared spectroscopy, a ham quality and corruption portable detection device was developed with automatic modeling function, and a ham shelf-life early warning system was designed based on the Internet of Things technology. Based on the developed detection device, the visible/nearinfrared spectra of 74 ham samples were collected. After different pretreatments, the ham color (L*, a*, b*), pH value, and volatile base were established respectively. Partial least squares (PLS) prediction model for nitrogen (total volatile basic nitrogen, TVB-N) and total colony (total viable count, TVC). On this basis, a competitive adaptive weighted sampling (CARS) algorithm was used to screen feature variables, and a simplified PLS model was established. The root mean square errors were 0.3530,0.2529, 0.0961, 0.0354, 0.3724mg/(100g), and 0.4398lgCFU/g, respectively. The results showed that the device can be used for the detection of ham quality spoilage parameters. At the same time, the TVC growth law during ham storage was studied, the TVC growth curve was fitted by the modified Gompertz equation, and the TVC growth kinetic model was established. The shelf life was established by the TVC spectrum prediction model and growth kinetic model prediction model. Alibaba cloud server and MySQL database were selected as server-side development tools, and based on the TCP/IP network communication protocol, the detection data of the device was transmitted to the server-side in real time, and then saved in the database. Finally, the data of shelf life was fed back to the client, and real-time early warning of ham spoilage can be realized by monitoring the shelf life. The experimental test showed that the system can be used for real-time early warning of ham shelf life.