Rockburst is an important accident scenario in deep underground engineering. Because there are numerous, complicated factors that lead to rockbursts, their forecasting is a difficult task, which, based on an engineering analogy and geological analysis, requires the use of clustering methods in rockburst forecasts. Because the environmental causes of rockbursts are complicated, this clustering problem makes for a complicated random optimization problem (that is also a fuzzy optimization problem) that cannot be solved in a satisfactory manner using traditional methods. To improve the computational efficiency and accuracy of the traditional ant colony clustering algorithm, an abstraction ant colony clustering algorithm using a data combination mechanism is proposed. Based on an analysis of rockburst sample data and using an engineering analogy thinking by the abstraction ant colony clustering algorithm, a new method for forecasting rockbursts in deep underground engineering is proposed. A set of common engineering examples are used to verify the new algorithm. The engineering applications prove that, compared with the traditional ant colony clustering algorithm and based on their similar computational difficulty and complexity, the abstraction ant colony clustering algorithm produces results that are not only more accurate but are also determined more efficiently. As the complexity of the problem increases, the algorithm’s computational efficiency increases. In other words, the more complicated the problem is, the more efficient the algorithm becomes. Thus, the abstraction ant colony clustering algorithm is well suited to large complicated engineering problems.