When processing data with complex spatial attributes, it is easy to fall into local optima, resulting in inaccurate outlier detection results. A spatial attribute data outlier detection method is proposed using the artificial bee colony algorithm in artificial intelligence technology to address this issue. Construct key value pairs of data blocks, calculate the fitness of the solution, use greedy selection method to identify the center points of each cluster, and achieve a set of outliers in spatial attribute data. Based on the determined center points of each cluster, classify the outlier attributes of the set space attribute data, calculate the detection radius, and identify outliers in the subset. Simulate the foraging behavior of bees, calculate the local reachable density of outliers, and obtain better outlier solutions through iterative updates. Calculate the outlier factor corresponding to the object and detect outlier points. The simulation results indicate that the proposed method achieves a detection accuracy of 90% when detecting outliers, providing a new and effective tool for spatial data analysis.