An empirical equation accurately predicting the distortion of a printed image on a plastic substrate was formulated. Experiments were performed to isolate parameters and evaluate each factor independently. An empirical equation that accounted for the tension, temperature and web speed was formulated. The substrate strain rate showed a linear dependency with the temperature and the tension, and increased exponentially with the web speed. Due to the differences between real image value and experimental error, a compensation factor was added to the empirical equation. In order to validate this empirical equation, we also performed experiments on substrates with different radial and longitudinal elongation rates. It was concluded that the elongation of the substrate does not affect the image strain rate. Experiments with a thicker substrate and a higher elongation rate confirmed that the image strain rate is more significantly affected by the thickness of the film than the elongation of the substrate. This empirical equation can be utilized to predict the image size at different conditions, and the registration accuracy can be improved by predicting the distortion during the processing.