Today's charging strategies of Li-ion batteries are often designed off-line and do not balance optimal fast charging and battery lifetime. By a lack of adaptation during operation, the off-line charging strategies do not allow aging-related changes of the battery properties to be taken into account. As such, to achieve safe operation, batteries and cells are typically operated in a range that is far away from the achievable operational limits. This article presents a nonlinear predictive control strategy that allows real-time optimal health-aware charging based on detailed first-principles models. The approach balances fast charging while improving battery lifetime. By considering the on-line adjustment of battery model parameters, the approach is able to directly take aging- and environment-related changes into account.