Abstract: Food safety has become a focus and the priority of people's livelihood. The issues of safety of agricultural products are frequent occurred around the world in recent years. For food safety incidents in the society, while agricultural products due to the presence of large uncertainties in producing area conditions, planting and processing, quality supervision of the entire journey quality traceability becomes extremely urgent for agricultural products. Through to the food safety supervision and research situation around the world, it can be found that in the construction of food safety supervision work in countries around the world,, government is gradually strengthening the construction of supervision mechanism and implementing mandatory measures to guarantee food safety. Different kinds of food safety traceability system are also gradually established and put into use. Agricultural products quality traceability system has become an important mean to solve the problem of food safety. But the current research is mainly on the single product traceability system. The public service platform has not yet been reported. The problems for lack of such system development mainly include the inability to support many kinds of traces, lower accuracy in the process of the trace, high cost of doing business application. In this paper, we developed a system applicable to quality traceability of a variety of commonly used agricultural products by analyzing the software framework of universal system and the traceability information data structure characteristics, focusing on the dynamic expansion of the data model optimization techniques and tracing precision lifting technique based on LBS. Aiming at the problem of narrow application range, we analyzed the data structure characteristics of traceability information index model, and compared four dynamic extended data models. By using the method of combining EAV with XML data model, the dynamic extension of different variety tracing information was realized through visual interface. It was found that the model compared to the other two models shortened time by 65% and query efficiency by 175% through performance analysis. For traceability information on producing environmental conditions, we analyzed the various types of available data sources and characteristics, formulated the smallest unit of space and realization method. We also used LBS technology to achieve the automatic acquisition of the origin of the environmental information. As such we found that the technique significantly improved the traceability system of width, depth, accuracy. Therefore it enhanced the overall system tracing precision and reduced the data collection cost. Through the database performance analysis, we found that the system query efficiency increased by 1.5 times, less CPU time. By traceability system general comparison, it indicated that the system can apply to many varieties of agricultural products with a good accuracy. We tested the system in livestock products quality and safety in Qinghai province. The results showed that the system had good performance and functional versatility, the system had obvious advantage in query efficiency and traceability accuracy, suitable for all kinds of agricultural products quality traceability.