Protein detection and identification at the single-molecule level are major challenges in many biotechnological fields. Solid-state nanopores have raised attention as label-free biosensors with high sensitivity. Here, we use solid-state nanopore sensing to discriminate two closely related proteins, α-thrombin and γ-thrombin. We show that aptamer functionalization improves protein discrimination thanks to a significant difference in the relative current blockade amplitude. To enhance discrimination, we postprocessed the signals using machine learning and training algorithms and we were able to reach an accuracy of 98.8% using seven features and ensemble methods.