How can I predict the true label for data with incomplete features based on the trained model with data with more features?

Suppose I have a model that was trained with a dataset that contains the features (f1, f2, f3, f4, f5, f6). However, my test dataset does not contain all features of the training dataset, but only (f1, f2, f3). How can I predict the true label of the entries of this test dataset without all features?