Abstract: Difference-in-differences (DID) is a statistical method specific to the nonequivalent group design (NEGD) and used to solve the problem of imbalance between the intervention group and control group in quasi-experimental study design. It can be used not only for dealing with the baseline difference between the two groups, but also for controlling the influence of confounding factors to effectively estimate the intervention effect. This paper summarizes the design principle and statistical method of DID, an example with a continuous variable as outcome (potential daily dose of formaldehyde, mg/d) is used to fit a general linear model, and the results are explained. With respect to the binary outcome, a logistic regression model is introduced to explain the statistical theory and parameters of DID. There are a bunch of quasi-experimental study designs with nonrandomized grouping in public health interventions. It is reasonable to handle this kind of data by using DID in order to achieve statistical effects similar to the randomized controlled trial (RCT). Difference-in-differences will be widely applied in the public health program evaluation with all-pervading technique of DID.