Fracture morphology of shales represents the main feature related to their brittleness, and plays an important role in affecting the complexity of fracture network in shale reservoirs. However, no brittleness index model based on fracture morphology has been developed so far. In the present research, a detailed analysis on morphological characteristics of shale fractures under triaxial compression has been performed, focusing on the Longmaxi Formation shale in the Weiyuan area, China. As a next step, the fractal theory is used to establish a method to characterize fracture complexity. As a next step, the correlation between rock mechanics parameters, mineral composition, stress-strain curve, dilatability, and shale fracture complexity is analyzed, determining the most influencing factors affecting brittleness, whose weight is analyzed by applying the gray correlation method. Furthermore, a novel brittleness evaluation method based on shale fracture morphology is established. Results show that Young's modulus, peak strain, and dilatancy angle of shales are characterized by the highest correlation values with fracture morphology (0.4745, 0.6302, and 0.579, respectively), while Poisson's ratio, mineral composition, peak, and residual stress, and strain have little correlation. The calculation and analysis of a shale horizontal well using the brittleness index calculation method in this paper shows that the established method has a good correlation with the number of events of MS after fracturing, the reconstruction volume, which provides technical support for the optimization of fracturing sweet spot in shale horizontal wells, and the evaluation of shale reservoir compressibility.