With the development of data-driven scientific research paradigm, the role of agricultural science data in science and technology innovation is becoming more and more prominent, and consequently the methodological and technological research on the analysis and mining and application of agricultural science data is also developing, around the analysis and mining of agricultural science data there are still data semantic silos serious, as well as the data mining tools are incomplete, mismatched and poor adaptability of the scenarios, such as the outstanding problems.In this paper, we designed the platform architecture, constructed the analysis and mining engine, loaded the typical and professional analysis and mining algorithm tools, formed the online analysis and mining platform for agricultural scientific data, including the data layer, the domain data analysis tool layer, the automated mining framework layer, the online analysis engine layer, and the user interface layer, and developed four functional modules, namely, the data management, the component management, the scenario management, and the mining analysis. The platform is equipped with application scenario management, online analysis, automated mining and other functions, breaking through the problem of poor connection of "data resources-analysis tools-application scenarios", forming an online analysis and mining application environment integrating data resources, analysis models, component tools, scenario analysis and standard processes, supporting the whole process of online analysis and mining of agricultural scientific data from "data aggregation - mining and analysis chain - online analysis - scenario application", and realizing the concurrent online interactive computation and analysis of ultra-large-scale data and different scenario analysis applications.