Abstract: Objectives The interaction between ice and the propeller causes significant differences in load characteristics between ice-region pods and open-water pods. Full-scale measurement data is a key support for the safety and reliability design of ice zone pod thrusters. To obtain full-scale measurement data for ice-region pods and investigate the correlation mechanism between load transient perturbations in ice conditions and ice-propeller contact events and interaction patterns. Methods A hypothesis for classifying ice-propeller interaction patterns is proposed based on theoretical analysis. Full-scale operational data of pod thrusters was collected during an Arctic research expedition of a PC3-class icebreaker, and the relationships among vessel motion, icebreaking operation modes, and pod load fluctuations were examined. A physically driven method for automatic identification and classification of ice-propeller contact events is presented. By comparing the fluctuation characteristics of pod load parameters within different classification event windows and integrating wavelet transform techniques, further feature identification is achieved. Results The results validated the initial hypothesis: ice-paddle interaction patterns can be categorized into four types——milling, crushing, impact, and clogging. During impact icebreaking operation model, the pod torque transient exhibits amplitude fluctuations reaching up to 160% of the rated torque, demanding the highest requirements for motor output control and over-torque capability. The greatest propulsion capacity is required under low-speed, heavy-load conditions in layered ice zones. Ice thickness, ship speed, and test location exhibit high correlation with pod load levels. Conclusions The study established the correlations between pod load fluctuation characteristics and ice conditions as well as ice-propeller interaction patterns. The measured data provides quantitative basis for propulsion limit load design of pods in ice-covered areas, and lays the foundation for assessing motor’s over-torque capacity and issuing navigation safety warnings.