A method was established to identify washed and sun-dried coffee comprehensively and accurately by screening differential components. High performance liquid chromatography(HPLC) of coffee samples were collected for multi-chemometrics analysis， and then the full raw signals and common peaks' area data that came from the two different processing types of coffee samples were divided into calibration set and test set respectively using the Kennard-Stone algorithm in the ratio of 7:3. Principal component analysis (PCA) and partial least squares-discriminant analysis (PLS-DA) models were established based on those data primitively. In the following work， the raw data were processed to optimize PLS-DA model by using the sole algorithm and combinatorial algorithm， such as variable importance in projection (VIP)， backward interval partial least squares (BiPLS)， successive projections algorithm (SPA)， uninformative variable elimination-partial least squares (UVE-PLS)， competitive adaptive reweighted sampling-partial least squares (CARS-PLS) separately， and BiPLS-VIP/CARS/SPA/UVE together. After that， the correlation coefficients ×100-1 of X-variables in the discriminant function of the optimum model were labeled on the chromatogram， which was used for screening quality markers between the two types of coffee， and the highest correlation component was qualitative analyzed. The results demonstrated that a qualified prediction model could not be established using the full raw signals to distinguish washed and sun-dried coffee， while the BiPLS algorithm at 90 intervals was the best optimum proposal among all kinds of algorithms， which had good discriminatory ability with an accuracy of 83.33%， a Q2 value of 0.813 and an root mean square error of cross validation (RMSEcv) value of 0.217. The qualitative results of the standards illustrated that chlorogenic acid is one of the important differential components of coffee with two different processing methods. Above all， compared to the traditional method of using the single or multiple compounds contents information as quality evaluation criteria， this identification method could dig out the differential components entirely and accurately， which providing a new thought with reference value for the product quality assessment comprehensive.