A series of new arylpropenamide derivatives containing different aryl groups were synthesized, characterized, and evaluated for their anti-hepatitis B virus (HBV) activities. A new high accuracy QSAR model of arylpropenamide was constructed based on a more completely activities data and calculation parameter. The 2D-QSAR equations, by using DFT and multiple linear regression analysis methods, revealed that higher value of thermal energy (TE) and lower entropy (S⊝) increase the anti-HBV activities of the arylpropenamide molecules. Predictive 3D-QSAR models were established by SYBYL multifit molecular alignment rule. The optimum models were all statistically significant with cross-validated and conventional coefficients, indicating that they were reliable enough for activity prediction.