Abstract: In order to detect soluble solids content(SSC) of apple conveniently and rapidly, a portable apple soluble solids content spectrometer was designed with ARM9 as controller, miniature spectrometer as spectrum detector, WinCE as embedded operating system. The total number of 80 apples was for the calibration(n=60) and prediction(n=20) sets. Different spectral pretreatment methods, including smoothing, standard normal variate(SNV) and multiplicative scatter correction(MSC) were performed. Uninformative variable elimination(UVE) and successive projections algorithm(SPA) were proposed to select effective wavelength variables. Partial least squares(PLS) models combined with whole wavelength and selected wavelengths were established to detect SSC of apple. The results indicated that the best model was developed using PLS with respect to the spectra in whole wavelength range. The correlation coefficient (Rp) and root mean square error of prediction(RMSEP) for prediction sets were 0.853, 0.534 for SSC. The model is reliable and the predicted result is effective. The meter can meet the requirement of quick measuring SSC of apple. The study can provide a reference for development of portable apple soluble solids content spectrometer.