Web access prediction is an important research direction in Web mining. Markov models are well-suited for predicting Web access. Although higher-order Markov models have good predictions result, these models have several limitations associated with high state-space complexity and reduced coverage. These affect the prediction performance deeply. A new Web access prediction model, Hybrid-order Tree-like Markov Model (HTMM), is proposed in this article. The technique intelligently merges two methods: a tree-like Markov model method that aggregates the access sequences by pattern matching and a hybrid-order method that combines varying order Markov models so that the resulting model has a low state complexity, improved prediction accuracy, and retains high coverage. Experiments confirm its usefulness. It's suitable for applications in E-business, such as Web prefetching, link prediction, and recommendation.