Recently, digital steganography has been a popular way used to hide information into digital data. Specifically, image steganography is the most used to conceal information in an image. Image steganogra phy can be performed through different techniques, algorithms, and tools. Besides, many adversaries use steganography to conceal mali cious payload in images, and then they extract it by a dynamic way on the victim’s device to do their malicious purposes, this what is called stegomalware. In this thesis, we propose an anti stego-malware approach that has the goal to destroy the stegomalware content in an image to prevent the predicted malicious effects, without leaving any visual degradation on the image. The concept of our approach is to add a lightweight and robust noising to the image, in addition, to add an invisible watermark to it. To implement our solution, we have used two steganography algorithms to embed a malicious payload into three images to obtain stego-images. Then, we have added the lightweight noising and the invisible watermark to the stego-images to obtain protected images. Finally, after testing it, the result obtained show that our solution can destroy the stegomalware content in image in a short time and without leaving any impact on the visual quality of the image.