Hall-Petch slope (k) is an important material parameter, while there is a great challenge to accurately predict the k value of magnesium alloys due to a high dependence of k on the material parameters, deformation history and testing conditions. The present study demonstrates that machine learning could provide opportunities to overcome this challenge. Two machine learning models, artificial neural network (ANN) and random forest (RF), were built and validated using 138 data. The results showed that increasing the training data set would enhance the prediction efficiency of both models. Comparing to the RF model, the ANN model showed higher accuracy. The correlations between individual attribute and k values were also discussed.