The correlation between pixels isn't completely removed, which degrade the performance of compression because the transform and quantization steps are bypassed in the lossless image compression with high efficiency video coding intra-prediction. To improve this situation, additional methods such as residual differential pulse code modulation is used in conjunction with the intra-prediction. The basic principle of residual differential pulse code modulation is based on the calculation of an enhanced residue of a pixel which is obtained with subtraction of the residue of left or top neighbor pixel from the residue of the pixel. In this study, a new residual differential pulse code modulation method, sampled-weighted residual differential pulse code modulation, considering an average of two or three neighboring pixels with weight is developed. A simple and fast algorithm is also proposed to determine the weight coefficients to be used in calculating the weighted average for the three neighboring pixels case. To obtain their compression performances the proposed sample-weighted residual differential pulse modulation methods are applied to various test images. The compression ratios of the compressed images are obtained.