There are three problems such as order preservation, stability and positive and negative in the grey incidence model applicable to panel data. To solve these problems, this paper has the following efforts. Firstly, the grey incidence coefficient is constructed based on the difference characteristics and exponential function, and the direction judgment factor is proposed to represent the positive and negative correlation. Secondly, the fluctuation adjustment factor is designed to describe the relative fluctuation degree between sequences. Thirdly, indicator weights are determined using the entropy method. A grey difference incidence model for panel data is proposed. In addition, antisymmetry and order-preserving property are defined in grey incidence analysis, and the new model is proved to have antisymmetry and order-preserving property. Finally, the model is applied to analyze the trend of air pollution in the Beijing-Tianjin-Hebei region from 2013 to 2021, which shows the practicability of the model. Through two sets of Monte Carlo experiments, the robustness and order-preserving property of the model are verified. The model is more suitable for the panel data with non-fixed object arrangement.