Abstract: The annual improvement in China’s ecological environment leads to a continuous increase in the scale of bird activities， and the risk and incidence of civil aviation bird strikes also increase. In addition to low-altitude resident birds around airports， migratory birds pose significant threats to civil aviation safety throughout the entire flight route owing to their wide spatial distribution range and large size. Reliable methods for monitoring wide-area bird situations are required for threat warning and reducing bird strike risks along the entire route. Meteorological radar systems， which have advantages like all-weather operation， large detection range， and networking capabilities， represent the best remote sensing solution for wide-area bird flock surveillance. In this study， we construct a high-dimensional feature space for bird flock targets based on meteorological radar detection data and uses data optimization and machine learning models to identify bird flock targets based on meteorological radar detection data. Based on this， a method for assessing bird flock density suitable for non-uniform spatial distribution conditions is proposed， enabling the quantitative description of the scale of bird flock activities within a specified spatiotemporal window. Based on prior information on the correlation patterns between bird activities and meteorological information， a model is constructed for mapping bird flock density with multiple meteorological parameters. By designing a reasonable data sampling method， a dataset containing bird flock density and meteorological parameters is generated， and a random forest regression model is used to establish a numerical regression model for bird flock density based on meteorological forecast information. Using meteorological radar detection data from Beijing and Fuzhou during the spring and autumn seasons， we validated the model from two aspects： cross-verification of radar and manual observation of bird situations. Experimental results showed that meteorological parameter information can effectively estimate bird flock density， and under high activity levels， bird flock density has higher prediction accuracy. This provides reliable reference information for evaluating wide-area bird strike risks and avoiding their flight routes.