The granular-ball (GB)-based classifier exhibits adaptability in creating coarse-grained information granules as input, thereby enhancing its generality and flexibility. Nevertheless, current GB-based classifiers rigidly assign a specific class label to each data instance and lack the necessary strategies to address uncertain instances. Such certain classification approaches to uncertain instances may suffer considerable risks. To solve this problem, We introduced the three-way decision into granular-ball SVM (GBSVM) to construct a robust three-way granular-ball SVM (3WGBSVM) model for uncertain data, which categorizes data instances into certain classes and uncertain cases. Extensive comparative experiments are conducted with 4 GB-based classifiers on 6 public benchmark datasets. The results show that our model demonstrates robustness in managing uncertain data and effectively mitigates classification risks. Furthermore, our model almost outperforms the other comparative methods in both effectiveness and efficiency.