A dynamic model of a batch crystallizer was used to estimate the kinetic parameters in a sugar cane crystallization process. The mathematical model included mass, energy, and population balance differential equations, and empirical relationships (power-law-type) for nucleation, growth, and production-reduction rates. Experimental data were obtained from runs carried out at different agitation and cooling profiles. A nonlinear optimization process was used to fit 11 kinetic parameters to experimental data. The overall experimental error was estimated in 7.71% and the average asymptotic error of all kinetic parameters was less than 2%. The nucleation rate was favored by high agitation speed and natural cooling profiles. The maximum growth rate was reached when supersaturation was at a maximum. There was a no clear effect of agitation profiles on growth. The overall effect of the term production-reduction was in all cases positive, with the largest values found for natural cooling profiles. (C) 2004 American Institute of Chemical Engineers.