Cyclin-dependent kinase 6 (CDK6) is a protein target for anticancer drugs. Its structural and binding affinity data made it possible to define the structural basis for its inhibition. This chapter focuses on building a neural network model to estimate binding affinity using the CDK6-pose complexes. Molegro Virtual Docker (MVD) generated the protein-pose complexes for small molecules for which binding data is available. It also determined ligand descriptors, energy terms, and scoring functions. They are the features employed to build regression models with Molegro Data Modeller (MDM). The regression model built here shows superior predictive performance compared with the Plant Score (a classical scoring function). The intelligent approach employed to build machine learning models emulates the exploration of the scoring function space. It constructs a model targeted to a protein of interest. The built neural network model can sort and estimate binding affinity for any docking screen using MVD. All CDK6 datasets and Jupyter Notebooks discussed in this work are available at GitHub: https://github.com/azevedolab/docking#readme .