Abstract: FWHM and image motion are two frequently used methods in the field of astronomy. However, estimation accuracy of these two methods will be affected by tracking error, wind and mechanical vibration. Raising its accuracy plays a rather important role in the area of high-resolution imaging, site selection and seeing monitoring of observatories, etc. For FWHM method, we fit the long exposure image with a two-dimensional gaussian function and choose an appropriate direction to measure its FWHM by which we estimate the seeing parameter. A comparison with spectral ratio method shows a great improvement of our approach against tracking error, wind and mechanical vibration. For image motion method, we employ principle component analysis to project center of gravity coordinates of a sequence of tracking error affected speckle images onto a proper direction on which we find its variance. We further utilize the variance to estimate the seeing parameter. A comparison with spectral ratio method also denotes a considerable improvement of our approach against tracking error, wind and mechanical vibration.