This paper presents a novel approach to developing a work roll prediction model that takes into account both the mechanism and condition influences on work roll wear. This was accomplished by conducting an analytic calculation of work roll mechanism influence, constructing a work roll wear model, and combining the wear mechanism with actual wear data. The resulting model is applicable to both symmetric and asymmetric wear of the work roll, and experimental results showed that the relative error between measured and predicted values was less than 5%, with a maximum error of below 15%. This level of accuracy is sufficient for predicting roll wear and lays the foundation for improved strip shape control and roll design. Furthermore, this approach has the potential to generate significant economic benefits and has wide-ranging applications.