To investigate the impact of environmental changes on the phenological stages of winter wheat, a predictive model for key phenological periods was developed. Environmental factors such as temperature, light hours, and rainfall were selected for analysis. Data characteristics were examined using scatter plots, normal distribution tests, and Pearson correlation analysis. Initial modelling was performed using multiple linear regression. However, due to the limited dataset and multicollinearity among independent variables, some parameter estimates failed significance tests. Therefore, the variance inflation factor (VIF) was employed to test the multicollinearity of the data, and six predictive models were tested for comparative evaluation. Among these, a model combining principal component analysis (PCA)with support vector regression (SVR) was identified as the most effective for phenological prediction. The results showed that the root mean square error (RMSE) for the verification of all phenological prediction models was below 1, with determination coefficients exceeding 90%, thereby demonstrating high predictive accuracy. This predictive model not only elucidates the influence of environmental factors on winter wheat phenology, but also  serves as a valuable tool for optimizing agricultural production, resource allocation, and risk management. Furthermore, it provides insights for agricultural climate suitability studies, scientific research and informed decision‑making in the agricultural sector.