The evolution of web and internet and data makes security at critical importance. More precisely the traffic related to image data. In this work, we suggest a novel image encryption algorithm based on chaotic map and a statistical indicator. In cryptographic context, an efficient encryption scheme should adopt the concepts of confusion and diffusion preconize by Claude Shannon. In confusion process, we propose a dynamic cyclic shift with respect to the statistical indicator mode combined with the XOR operator. In diffusion mechanism, we adopt a special process that relate each pixel with neighborhoods pixels. Our method is evaluated using histogram, entropy analysis, correlation coefficient, Number of Pixels Change Rate (NPCR), Unified Averaged Changed Intensity (UACI) and key space. The experimental tests show efficient results with this algorithm. Indeed, the key space is large enough, which allow our encryption scheme safer against brute force attacks, the entropy is closed to the theoretical value (8), the encryption image histogram is flat, the NPCR and UACI values reflect that the proposed method is resilient against differential attacks.