Predicting the pixel is crucial in prediction error expansion (PEE) based reversible data hiding (RDH). There are a number of methods for predicting pixels that can be found in research papers. The gradients are used to make a prediction about the current pixel. Gradients in the image can be used to make forecasts about the current pixel, and this has been the subject of extensive study. In this work, we suggest a three-by-three grid of surrounding pixels as a basis for prediction. Later, the histogram bin shifting technique was implemented to integrate more information with less distortion. In order to include more information, the suggested technique makes adaptive changes to the histogram based on local complexity. The experimental analysis demonstrates that When compared to the other, the proposed method is superior.