This paper develops an optimum replacement model, in the context of Condition Based Maintenance (CBM), in order to minimize the long-term average cost of equipment, using Logical Analysis of Data (LAD). LAD has the advantage of not relying on any statistical theory, which enables it to overcome the conventional problems concerning the statistical properties of the datasets. LAD’s main advantage is its straightforward procedure and self-explanatory results. In this paper, our main objective is to develop a method to minimize the long-term average cost of equipment maintenance, using LAD. We employ LAD’s pattern generation procedure. Using the generated patterns, we estimate the equipment’s survival probability. These probabilities are then used in a dynamic programming model to make optimum decision about when to replace a piece of equipment under condition monitoring. The proposed methods are applied on Prognostics and Health Management Challenge dataset, a condition monitoring dataset collected from mechanical equipment provided by NASA Ames Prognostics Data Repository. Analysis of performance of the proposed methods reveals that the methods provide reliable results that are greatly beneficial to maintenance practitioners. Optimum replacement cost obtained by the proposed methods are compared with that of Proportional Hazards Model (PHM).