Depending on near infrared absorp tion spectrum of textile fiber between 1300nm and 1800nm, near infrared (N IR) spectrum technology is used to predict the chemical properties of textile fiber, coup led withmultivariate analytic statistical techniques. In this paper, we adopt stepwise multip le linear regression ( SMLR) to select characteristic wavelength of textile fiber, and to research the concentration of the mixture of cotton and terylene. By computing corre- lation coefficient,we establish calibration equation with the smoothed absorbance data. The error is controlled under6%. The experiment shows thatN IR technology can be used as amethod of qualitative analysis and quantitative analy-sis for textile fiber.