Geochemical data evaluation is a key research content in oil and gas exploration and development. In the evaluation process, it is often necessary to comprehensively use various discrimination diagrams to analyze data in multiple dimensions. With the accumulation of data and the increasing demand for detailed analysis, the challenges faced by qualitative and quantitative evaluation of geochemical data are growing day by day. This paper proposes an intelligent evaluation method for geochemical data based on the scatter data fence model. By using the classification lines and discrimination diagram axes of the discrimination diagram, the discrimination diagram is divided into several closed sub-zones. Several characteristic scatter points are selected on the edge of each closed sub-zone to establish a polygon data fence model. The ray method is used to detect the inclusiveness between geochemical data points and polygon data fence models, and the program automatically determines the projected position of geochemical data on the discrimination diagram, thereby achieving accurate evaluation of the data and quantitative statistical analysis. This method effectively solves the problems of low accuracy of traditional data mapping and machine learning method evaluation results, complex model training processes, and difficulty in quantitative statistical analysis. It can evaluate data more efficiently, shorten the evaluation period, save manpower resources, and meet the needs of accurate evaluation and quantitative statistical analysis of basic data in the background of big data analysis.