Thermal properties such as melting and crystallization are important aspects in understanding the morphology and its contribution to the physical properties of semicrystalline polymers, such as polypropylene. The inclusion of fillers, which are small particles dispersed in the continuous polymer phase, often complicates the predictability of these properties by acting as nucleating agents or defect origins. This paper discusses the creation and use of empirical models based on experimental data for predicting and optimizing the thermal properties of agricultural filler-polypropylene (AgFiller-PP) composites, including peak melting temperature (Tm), peak crystallization temperature (Tc) and percent of crystallinity (Xc). Experiments were performed using differential scanning calorimetry (DSC) to gather data necessary for building appropriate prediction models. Finally, additional experiments were carried out to test the prediction results generated by the models.