Abstarct: As the oil and gas exploration becomes gradually complicated, the traditional well logging method has many problems. With limited information, recognizing effective laxyers and estimating reserves parameters become more and more difficult. In contrast, the Formation Micro Imager (FMI) technology can provide rich information on fractured reservoirs, and most importantly it can be applied to identify fractured reservoirs qualitatively and can help explain them quantitatively. Iranian reservoirs are mainly carbonate reservoirs in which the fractures have important effect on permeability and petroleum production. . To our knowledge, we rarely find automatic fracture recognition systems with FMI images in the literatures. Consequently, for promoting oil and gas exploration, it is of great significance to develop an efficient and intelligent recognition system using FMI images. The projection of a planar feature on borehole images will be represented by a sine wave. Because the borehole is very small in relation to Stratigraphic and structural features, when they intersect the borehole, they are nearly planar and assumed to be planar feature, so in this paper we use sine function for fitting the best curve on fracture pattern. We analyze rock FMI images with very limited samples in a different way by using image processing. There are several key issues in such recognition systems for FMI images. Firstly, the training samples are limited in most cases due to sensitive company data source. Secondly, FMI images are usually of great noises. In practice, even with manual processing, it is very difficult to analyze and classify rock structures without enough expertise. So, some fractures cannot be detected using this algorithm because of great noisy FMI images and also existence of other features and rock textures with low resistivity around fractures which cause the ambiguous condition in processing. Moreover, related references were very limited and we used just few references. The dip and azimuth of fractures are detected using this algorithm to identify more precise permeability and career of fluid in reservoir. Our algorithm includes three main steps. First, pixels represented fracture are extracted from projected FMI image involve location matrices x and y and corresponding values matrix f(x,y). Then, two vectors X and Y as inputs of CFTOOL of MATLAB are produced by combination of these three matrices. Finally, optimum combination of sine function is fitted to sine shape of fracture pattern to identify dip and azimuth of fracture. We test our fracture recognition system with real interpretation FMI rock images. In experiments, the average recognition error of our proposed system is about 0.7 % for azimuth detection and less than 3% for dip. The results are improved by statistical analysis so that the confidence level for correlation between real and determined parameters is upgraded to 0.95 (95%). Moreover, this automatic system can significantly reduce the complexity and difficulty in the fracture detection analysis task for the oil and gas exploration.