Optimization of the manufacturing process based on scientific evidence is essential for quality control of active pharmaceutical ingredients. Real-time monitoring can ensure the production of stable quality crystals in the crystallization process. Raman spectroscopy is an attractive tool for pharmaceutical quality evaluation and process analytical technology because of its ability to analyze samples non-destructively and rapidly. In this study, we attempted to monitor the crystal polymorphs of carbamazepine (CBZ I and CBZ III) during the dissolution and crystallization processes using low-frequency Raman spectroscopy, which can reflect differences in lattice vibrations originating from polymorphs in the scattering peaks. Furthermore, using multivariate analysis of the obtained spectra, we attempted to develop a model that enables the quantification of each polymorph. A partial least squares was performed to build the prediction model. The prediction model was built using a set of 33 calibration samples, and an external set of 12 validation samples was used to evaluate the model. The model presents a good prediction capacity. The quantitative results for the solid amount of carbamazepine in suspension calculated using the model during the dissolution and crystallization process showed results that correlated very well with the particle view results. It is suggested that low-frequency Raman spectroscopy can be used as a useful process analytical technology tool.