The use of the posterior Cramér-Rao lower bound (PCRLB) as a lower bound for the mean-squared estimation error (MSEE) of progressive damage is investigated. The estimation problem is formulated in terms of a stochastic dynamic system model that describes the random evolution of damage and provides measurement uncertainty. Based on whether the system is linear or nonlinear, sequential Monte Carlo techniques are used to approximate the posterior probability density function and thus obtain the damage state estimate. The resulting MSEE is compared to the lower bound offered by the PCRLB that is obtained from the implied state transition probability density function and the measurement likelihood function. The progressive estimation results and the PCRLB are demonstrated for fatigue crack estimation in an aluminum compact-tension (CT) sample subjected to variable-amplitude loading.