Abstract: In accordance with the characterics of the large number and complex and unpredicted trend on change of the meteorological detecting data,the piecewise curve fitting method based on the least squares of orthogonal polynomial for scattered data is presented.Firstly,the reason of using the least squares of orthogonal polynomial for scattered data is introduced from the respect of the stability of data processing.Secondly,the principle of twice piecewise curve fitting is set out in detail from the respect of the shape preserving feature of curve fitting.Finally,as the temperature detecting data for example,it is proved clearly that the twice piecewise curve fitting method can make much better fitting precision than the traditional interpolation method,and improve the level of precision and automation of the meteorological detecting data processing.

Key words: scattered data, orthogonal polynomial, least squares, piecewise curve fitting, meteorological detecting