Recently, RGB-White(RGBW) sensors have been developed to take advantages of high-sensitive white(W) channel. Especially in the low light condition, the W channel has better SNR (Signal to Noise Ratio) characteristics than other color channels. In this paper, a new color interpolation method for the RGBW sensor is proposed. To make the best use of the high-sensitive W channel, we use a random pattern RGBW Color Filter Array(CFA). The most of pixels in the CFA sample the W channel values, when a small number of pixels sample the R, G, and B channel values respectively. At the initial step, missing W channel values in positions of pixels which sample the R, G, and B channel are estimated by edge adaptive scheme. After that, the color difference channels, i.e., R-W, G-W, and B-W, are estimated by the scheme of colorization. Experimental results show that the proposed algorithm performs better than the results from Bayer CFA in terms of SNR and object discrimination.