In the context of servitization in manufacturing, selecting equipment maintenance strategies involves significant complexity and uncertainty, making traditional methods often inadequate, which in turn affects maintenance costs and equipment availability. Currently, the multi-criteria evaluation information in maintenance decision-making is often characterized by fuzziness and uncertainty, making it challenging for traditional approaches to provide a scientific and well-grounded maintenance plan. To address this issue, this paper introduces an analysis model based on probabilistic linguistic term sets(PLTS) and proposes three weighted power average operators based on the PLTS to integrate multi-criteria evaluation information under both known and unknown criteria weights, effectively managing uncertainty and fuzziness in maintenance decisions. By applying the proposed methods, their feasibility and effectiveness in selecting maintenance strategies for servitized manufacturing supply chains are demonstrated, providing a more scientific tool for decision-making that helps improve cost control and enhance equipment stability.