Designing new watermarking models for the authentication of transmitted images over wireless multimedia sensor networks is an essential need. Indeed, wireless networks require considering several constraints including execution time, robustness, and fault tolerance against different attacks. Computational intelligence and knowledge discovery approaches, such as rough set and Formal Concept Analysis (FCA), make it possible to efficiently exploit the host image characteristics to identify interesting locations for watermark embedding. In this paper a robust watermarking approach based on FCA is proposed. It derives relationships between the image’s blocks described through a set of texture features, and uses these relationships to identify the more interesting blocks for watermark embedding. The experiment’s result proves the efficiency of the proposed technique in terms of imperceptibility and robustness against different attacks.