To make realistic maintenance decisions, it is important that maintenance managers make their preventive replacement decisions based on observations of the condition of their equipment. This study addresses a condition- and age-based replacement decision problem using the complete history of measured condition observations to minimize long-run average cost, maximize long-run average availability, or both. A stochastic filtering process (SFP) is used to estimate the residual lifetime distribution conditional on the history of observed condition information. A long-run average cost model and a long-run average availability model are analysed in order to determine the theorems necessary for calculating the optimum replacement time. To minimize the cost and maximize availability, a multiobjective decision frontier is proposed that will help maintenance managers deal with trade-offs between the two objectives. Finally, numerical examples are presented for each scenario to show the effectiveness of the methods proposed.