The inversion of pollutant emissions is important in air pollution prediction and control. The statistical methods generally adopted possess weaknesses, such as sensitivity for observation error and prior estimation of emissions. Given their simplicity, high precision, and practicality, algorithms based on the shooting method are widely used in the field of systems control. In this paper, we derive a shooting method-based adaptive algorithm to estimate air pollutant emissions. Besides its high precision and simple procedure, this adaptive algorithm compensates for the weaknesses of statistical methods. It is able to deal with the large level of error in the emissions inventory, initial conditions, and observations; a prior distribution assumption and error estimation are not required. Simulations of a simple system are presented to illustrate the effectiveness of this adaptive algorithm.