Star clusters, even unresolved, are ideal chronometers for the study of the star formation history of normal and starburst galaxies in the local Universe. Star cluster ages and masses can be derived from their spectra/photometry by comparing these with cluster evolution models. However, presently available spectrophotometric cluster evolution models do not take into account cluster disruption, mass segregation and the preferential ejection of low-mass stars from the clusters and the resulting changes in the clusters' stellar mass function, effects both seen and of great importance in observations and dynamical cluster models. This leads to (potentially large) errors in cluster age determinations and in the derived star formation histories of galaxies if "standard" approaches are taken. We will study the disruption of clusters and changes in their mass function in various environments, using N-body simulations including mass segregation and binary stars.This will be combined with a n improved version of a state-of-the-art cluster evolution program to produce the next generation of spectrophotometric cluster models (spectra, magnitudes, colours, spectral indices) as a function of the physical parameters of the cluster (age, mass, metallicity) and the disruption time scale (intrinsic host galaxy's property).