Many enzymes have been widely used in industrial production, for they have higher catalytic efficiency and catalytic specificity than the traditional catalysts. Therefore, the performance of enzymes has attracted wide attention. However, due to various factors, enzymes often cannot show their greatest catalytic efficiency and the strongest catalytic ability in industrial production. In order to improve the enzyme activity and specificity, people become increasingly interested in the transformation and modification of existing enzymes. For the structure modification of proteinase, this chapter introduces a computational method for modelling error-prone PCR. Error-prone PCR is a DNA replication process that intentionally introduces copying errors by imposing mutagenic reaction condition. We then conclude about the mathematical principle of error-prone PCR which may be applied to the quantitative analysis of directed evolution in future studies.