The mineral constituents of the rock sample can be analyzed with in-situ energy dispersive X-ray fluorescence analysis technology (In-situ EDXRF), the matrix effect of rock sample will effects on measurement results. The Monte Carlo simulation method is used to conduct fluorescence analysis spectrum with ideal measurement conditions, which provides analytical data for matrix effect research. The measured spectrum of seventeen kinds rock samples are being simulated, which has the same Cu content. Therefore, the influences with matrix effect of rock sample in in-situ EDXRF take Cu element for example. Based on correlation between Cu Kα X-ray intensity and spectral parameters, considering elements similarity of all kinds rock samples, it is found that the variation the Cu Kα X-ray intensity not only by the control of rock elements composition or rock classification. The matrix effect of rock samples must be classified according correlation between Cu Kα X-ray intensity and spectral parameters. After the matrix effect classification, fifteen kinds of rock samples, which belong to the same matrix effect, can be corrected more effective. Based on principal component analysis of similar matrix effect rock samples, it is found that the scattering background, target element K-series X-ray of X-ray tube and its incoherent scatter intensity can be a good description of Cu Kα X-ray intensity which is affected by rock matrix, thus it can be used to correct the Cu element measurement results. Certainly, this technology can also provide reference for matrix effect correction to other elements in rock.