Patchy vegetation pattern is a common type of landscape in arid and semi-arid regions of the world. It is of great significance to study their formation, structure, and succession processes for revealing the key processes underlying the dynamics of an ecosystem. It is also helpful in developing suitable restoration strategies for degraded ecosystems and to cope with the impacts of climate change in arid and semi-arid regions. Considering that it is difficult to comprehensively elaborate the processes and mechanisms of vegetation pattern formation based on ground investigation and remote sensing technology, model simulation is an effective method to deal with this problem. Since the beginning of 1990s, a continuous and discrete modeling of patchy vegetation pattern formation has been emerging. However, the comparison and validation of continuous models to the real vegetation patterns in nature, and the complexity of discrete models to the rule-making still need to be studied. According to a brief review of the positive and negative feedbacks underlying pattern formation, we have focused on the new advances in continuous and discrete modeling of patchy vegetation pattern formation in this study, and also pointed out the shortcomings of previous research. The feedbacks between plant and soil water, operating at small scales, lead to a large-scale patchy vegetation pattern. A complete understanding of the mechanism involved in the feedbacks between plant and soil water is key to modeling patchy vegetation pattern formation. An external environment such as grazing intensity and precipitation pattern has an important influence on the characteristics of patch vegetation pattern in arid and semi-arid regions. Further, we have proposed a guide on strengthening the comparisons of the model results with observation-based patterns, parameterization, and validation data, paying attention to the inclusion of local environmental conditions and ecosystem functions in the model, building a hybrid model based on the advantages of continuous and discrete models, developing standard sub-models for key processes and software platforms, and emphasizing pattern-oriented modeling and spatial explicit modeling of vegetation pattern formation.