Abstract: New quality productivity has gained wide attention for its main features such as, high technology, high efficiency and high quality. Among them, vigorously developing the digital economy and promoting the integration of digital and real are the focuses of new quality productivity. And the data element is constantly exploring progress as the fifth production factor as well as the basis for realizing the digital intelligence of the manufacturing industry. A method of data reliability evaluation of digital twin machine tools is proposed. The node diagram of data interaction network is constructed by simulation system, and the weight of data interaction between systems is calculated based on data transmission frequency, data throughput, data delay and data importance. Theil's inequality coefficient (TIC) method was used to calculate the reliability of data, and the combined reliability threshold and the reliability threshold calculated by a single TIC method were calculated, and the validity of this method was verified..