Optomechanical crystal cavities are devices based on optomechanical interactions to manipulate photons and phonons on periodic subwavelength structures, enabling precise measurement of the force and displacement. The performance of the target structures varies when applied to different applications. Optomechanical crystal cavities now rely on an empirical forward design, which is inefficient. Therefore, a desired shift is toward directed design with a “problem-oriented” strategy. The directed optimization problem’s nonconvex nature and extensive parameter space necessitate substantial computational resources, driving the need for intelligent algorithms in a sub-wavelength structure design. Intelligent algorithms can surpass the constraints of traditional methods and discover novel structures that are effective in different materials, topologies, modes, and wavelengths. This paper provides an extensive overview of intelligent algorithms for guiding the directed design of optomechanical crystal cavities. It presents a systematic classification of 15 algorithmics, including, but not limited to, topology algorithms, particle swarm optimization algorithms, convolutional neural networks, and generative adversarial networks. The article provides a comprehensive review and thorough analysis of the principle and current application state, as well as the advantages and disadvantages of each intelligent algorithm. By using these intelligent algorithms, researchers can enhance the efficiency and accuracy of optimizing optomechanical crystal cavities in a broader design space.