This study assessed factors related to pressure ulcer development in patients hospitalized at an acute care university hospital with advanced pressure ulcer measures. Analysis was conducted with an alternating decision tree (ADT) using data in the hospital database. Assessed data included sex, age, height, weight, disease, surgery/no surgery, operation duration, nursing needs score A/B items, and data related to pressure ulcers. Based on a 5-stump analysis of the pressure ulcer and non-pressure ulcer groups, operation duration, sex, transfer activity, and BMI were identified as factors related to the development of pressure ulcers. The tree diagram had an accuracy, sensitivity, and specificity of 73.49%±4.26, 82.00%± 18.87, and 73.44%±4.34, respectively. The current dataset was imbalanced, with 51 patients in the pressure ulcer group and 8235 in the non-pressure ulcer group. When using such imbalanced data, there is the drawback of generating a complex decision tree. However, ADT has the strength of generating a single highly accurate tree by combining stumps with high reliability. Unlike other factors, transfer activity, which was extracted as a factor and is a nursing needs score B item, is amenable to nursing intervention. Our findings suggest that pressure ulcers can be predicted by focusing on the degree of independence with respect to transfer activity at the time of hospitalization, based on the daily evaluation of nursing needs data and use of data obtained at the time of hospitalization.