The problem addressed in this paper is disassembly sequence planning for the purposes of maintenance or component upgrading, which is an integral part of the remanufacturing process. This involves disassembly, component repair or replacement, and reassembly. Each of these steps incurs cost as well as the probability of damage during the process. This paper presents a method for addressing these tradeoffs, as well as the uncertainty associated with them. A procedure for identifying the best sequence of disassembly operations for maintenance and/or component upgrade is presented. It considers both disassembly and reassembly costs and uncertainties. Graph-based integer linear programming combined with multiattribute utility analysis is employed to identify the best set of tradeoffs among (a) disassembly time (and resulting cost) under uncertainty, (b) the probability of not incurring damage during disassembly, (c) reassembly time (and resulting cost) and (d) the probability of not incurring damage during reassembly. An example of a solar heating system is used to illustrate the method.