The quantitative structure-activity relationship (QSAR) of 30 acylthiourea analogues was studied by using a three-dimensional holographic vector of atomic interaction field (3D-HoVAIF) to describe their chemical structures. The descriptors obtained were screened by stepwise multiple regression (SMR) and a partial least-squares (PLS) regression model was built. The correlation coefficient r2 of the established model and Leave-One-Out (LOP) Cross-Validation (CV) correlation coefficient q2 are 0.624 and 0.409, respectively. The model has favorable stability and good prediction capability, and further QSAR analysis showed that hydrophobic interaction has the most important effect on the activity of acylthiourea analogue and 3D-HoVAIF was applicable to the molecular structural characterization and biologicalactivity prediction.