: This paper proposes a novel statistical evaluation model for automated protein NMR sequential resonance assignment. It can be bound to any assignment program and provides confidences for the whole output assignment and each individual mapping. A simulation study on a set of four proteins shows that the statistical evaluation results are informative.