Crystallization is a major challenge in metallic glass production, and predictive models may aid the development of controlled microstructures. This work describes a modeling strategy of nucleation, growth and the dissolution of crystals in a multicomponent glass-forming system. The numerical model is based on classical nucleation theory in combination with a multicomponent diffusion-controlled growth model that is valid for high supersaturation. The required thermodynamic properties are obtained by coupling the model to a CALPHAD database using the Al-Cu-Zr system as a demonstrator. The crystallization of intermetallic (Al, Cu)(m)Zr-n phases from the under-cooled liquid phase were simulated under isothermal as well as rapid heating and cooling conditions (10(-1)-10(6) Ks(-1)). The obtained time-temperature transformation and continuous-heating/cooling transformation diagrams agree satisfactorily with the experimental data over a wide temperature range, thereby, demonstrating the predictability of the modeling approach. A comparison of the simulation results and experimental data is discussed.