In order to improve the screening efficiency of domestic civil aircraft performance monitoring parameters, a method based on improved sliding time window is proposed to screen stable cruise parameters. Unscented Kalman filter (UKF) introduces sage husa noise estimator, constructs adaptive unscented Kalman filter (AUKF), and uses AUKF to denoise QAR data; The selection standard of cruise parameters is determined, the recursive algorithm improves the sliding time window algorithm and realizes the selection of cruise parameters, and the GUI develops the parameter selection system to further improve the efficiency of parameter selection. Through the sample data of domestic ARJ21 aircraft, the results showed that AUKF can improve the reliability of data, and the improved sliding time window algorithm can improve the screening efficiency by about 50%. The research can provide reliable and efficient data selection algorithm basis for domestic civil aircraft performance monitoring parameter selection.