Agricultural product network marketing still has problems such as weak infrastructure, incomplete logistics system, and unremarkable product brand, and there is a little combination of basic theory and practice. In order to improve the marketing effect of agricultural products, based on the intelligent recommendation algorithm, this paper combines data mining technology to construct an intelligent agricultural product business and product marketing recommendation model. Moreover, from the perspective of users and commodities, this paper proposes a general method of reordering recommendation results. In addition, this paper understands the results from the user perspective and the results from the commodity perspective as row sorting and column sorting respectively, and designs a linear fusion method of row sorting matrix and column sorting matrix to obtain new recommendation results. Finally, this paper constructs a framework model for agricultural product promotion, and analyses the performance of the model constructed in this paper from an experimental perspective. On the whole, the agricultural product marketing model constructed in this paper is effective.