Standard particle filter can solve the problem of maneuvering target tracking effectively under glint noise, but severe degradation phenomenon sometimes exists in sampling process. For this, an improved particle filtering algorithm was presented; degradation detection was carried out in sampling process, if the particles are seriously degraded, re-sampling until degradation within the allowable range. Simulation results show that under the Gaussian environment, the improved particle filter has the same performance with that of standard particle filter, being superior to unscented Kalman filter, but under glint noise environment, its performance is significantly higher than that of unscented Kalman filter and standard particle filter, but consuming longer work time.