Abstract: In order to explore the vacuum drying characteristics of yam slices, a drying model was established. Vacuum drying tests were carried out under different drying temperature (40, 50, 60, 70, 80 ℃), vacuum degree (0.03, 0.04, 0.05, 0.06, 0.07 MPa) and slice thickness (2, 4, 6, 8, 10 mm). BP neural network model and Weibull distribution function were used to fit the experimental data, and the effective moisture diffusivity coefficient, drying activation energy were calculated. The results showed that the temperature (P=7.56×10⁻¹¹) and slice thickness (P=1.82×10⁻⁶) had significant effect on the drying time, but the vacuum degree (P=0.32) had no significant effect on it. The average relative error of BP neural network model was 3.08%, which was lower than that of Weibull distribution function by 10.7%. BP neural network was more suitable to describe the vacuum drying process of yam slices. The effective moisture diffusivity coefficients ranged from 4.0042×10⁻⁹ to 3.4652×10⁻⁸ m²/s, which was greatly affected by temperature and slice thickness. The drying activation energy was 33.802 kJ/mol. This study would provide theoretical basis for vacuum drying of yam slices.