A new algorithmic procedure for resolving superimposed spectra and kinetic data of transients produced in fast complex chemical processes is presented. The optimization model allows partitioning of the unknown model parameters into two groups: linear spectroscopic parameters and non-linear kinetic parameters. When solving the inverse problem, the field of iterations is confined to the kinetic parameters and so no initial information regarding spectroscopic behaviour of the transients is required and the computing time is drastically reduced, as compared with a previous method. We illustrate the optimization model for the reaction of multisite hydrogen attack on three dyes (acridine, proflavin and acridine orange) studied by pulse radiolysis. The reliability of the optimized parameters (k, ε) is critically examined by means of a sensitivity matrix. The optimization procedure, combined with pulse radiolysis, laser or flash photolysis, will be helpful in providing an insight into the reaction mechanisms of fast complex chemical processes.