As an important source of small molecule drugs, natural products show remarkable biological activities with their rich types and unique structures. However, due to the limited number of samples and structural complexity, the rapid discovery of lead compounds is limited. Therefore, in this study, natural inhibitors of phosphodiesterase 4 (PDE4) and Phosphodiesterase 7 (PDE7) were screened by combining computer aided drug design (CADD) technology and deep learning method, and their activities were verified by enzyme activity experiment and enzymo-linked immunoassay. These two enzymes have important application potential in the treatment of inflammatory diseases such as chronic obstructive pulmonary disease and asthma, but PDE4 inhibitors may cause adverse reactions, so it is particularly important to develop both effective and safe dual-target inhibitors. In addition, as a potential target of hyperuricemia, the development of natural inhibitors of xanthine oxidase (X0) is also of great value. We used pharmacophore technology for virtual screening, combined with molecular docking technology to improve accuracy, and finally selected 16 potential natural inhibitors of PDE4/7, and verified their binding stability through molecular dynamics simulation. The results of this study laid a foundation for establishing an efficient dual-target inhibitor screening system and exploring the lead compounds of novel X0 inhibitors.