Pharmaceutical crystallization has an impact on the properties of active pharmaceutical ingredients as it determines the purity, crystal size distribution and polymorphs, among other attributes. Focused beam reflectance measurement (FBRM) is a common process analytical technology (PAT) tool used for monitoring, predicting, and control of crystal size and size distribution (CSD) characteristics during crystallization processes. Models for the prediction of crystal size distribution from FBRM statistics (i.e. chord length distribution, CLD) have been widely developed in academia and pharmaceutical industries but limited work has been done on the verification and validation of CLD to CSD models. This work demonstrates the development and validation of two CLD–CSD models, theoretical and empirical, for a model compound. The experiments were based on the cooling crystallization of carbamazepine (CBZ). Also, the precision and robustness of the FBRM statistics, square weighted mean chord length (SWMCL) and total counts, were evaluated. Agitation speed and solid concentration had an impact on the FBRM statistics. The theoretical geometric model evaluated in this study showed high error for the prediction of the CSD and it is not going to be implemented in on-going research efforts in continuous crystallization of CBZ. The data driven model can be implemented as a qualitative monitoring tool in continuous operation. The accuracy and robustness of the data driven model proposed need to be improved to be considered for quantitative monitoring of the CSD.