This study presents a quantitative structure–permittivity relationship (QSPR) model for predicting the dielectric constants (ε) of a diverse set of polymers. A transparent mechanistic model was developed using a machine learning approach that combines genetic algorithm and multiple linear regression analysis. The model was validated using various criteria, and four- and eight-variable models were proposed. The best model showed high predictive performance with R² values of 0.905 for training and 0.812 for test sets. The models were found to be robust and capable of predicting polymer permittivity accurately. The models were developed using 71 polymers with diverse structures, and the molecular descriptors were calculated based on the structure of repeating monomer units. The models were validated using leave-one-out cross-validation, y-scrambling, and internal/external validation protocols. The four-variable model showed good performance with R² values of 0.842 and 0.715 for training and test sets, respectively. The eight-variable model showed better performance with R² values of 0.905 and 0.812 for training and test sets, respectively. The models were found to be robust and suitable for further development of polymers with desired dielectric constants based on chemical structure information. The models were validated using a y-randomization test, and the results showed that the models were not due to random correlations. The models were also found to be suitable for predicting the permittivity of homopolymers. The study highlights the importance of molecular descriptors in predicting polymer properties and the potential of QSPR models in the development of new polymeric materials. The results of this study pave the way for future steps in investigating the electrical conductivity mechanism of polymeric materials.This study presents a quantitative structure–permittivity relationship (QSPR) model for predicting the dielectric constants (ε) of a diverse set of polymers. A transparent mechanistic model was developed using a machine learning approach that combines genetic algorithm and multiple linear regression analysis. The model was validated using various criteria, and four- and eight-variable models were proposed. The best model showed high predictive performance with R² values of 0.905 for training and 0.812 for test sets. The models were found to be robust and capable of predicting polymer permittivity accurately. The models were developed using 71 polymers with diverse structures, and the molecular descriptors were calculated based on the structure of repeating monomer units. The models were validated using leave-one-out cross-validation, y-scrambling, and internal/external validation protocols. The four-variable model showed good performance with R² values of 0.842 and 0.715 for training and test sets, respectively. The eight-variable model showed better performance with R² values of 0.905 and 0.812 for training and test sets, respectively. The models were found to be robust and suitable for further development of polymers with desired dielectric constants based on chemical structure information. The models were validated using a y-randomization test, and the results showed that the models were not due to random correlations. The models were also found to be suitable for predicting the permittivity of homopolymers. The study highlights the importance of molecular descriptors in predicting polymer properties and the potential of QSPR models in the development of new polymeric materials. The results of this study pave the way for future steps in investigating the electrical conductivity mechanism of polymeric materials.