The invention discloses a method for predicting the binding free energy of protein and ligand based on a progressive neural network, and belongs to the technical field of computer-aided drug design. The method comprises the steps: obtaining a pdb file from a PDBbind database, establishing local database, acquiring an amino acid molecule within 4.5 angstroms in the protein binding pocket by takingthe ligand molecule as a center, performing extended connectivity fingerprint calculation, carrying out SPLIF fingerprint calculation, searching for the number of salt bridges and hydrogen bonds between protein and ligand molecules, converting the structural information of the protein and the ligand into a one-dimensional tensor, and establishing a training set, a verification set and a test set;training the progressive neural network by using the training set; optimizing and searching hyper-parameters for prediction; through comparison with a molecular docking result, obtaining a higher Pearson correlation coefficient. According to the invention, the technical problem of how to convert a three-dimensional structure of protein and ligand molecules into tensors which are easy to calculateby a computer and input the tensors into the progressive neural network for training and optimization is solved, and the calculation rate and the prediction accuracy are greatly improved.