PurposeThe purpose of this paper is to determine the state probabilities for a float system when overhauls are performed periodically.Design/methodology/approachFirst, a model is developed to express the relationship between the rate of occurrence of failure and time interval between overhauls. Then, a methodology is defined to develop differential equations for each system state in order to calculate corresponding system state probabilities.FindingsSince it is considered that units requiring an overhaul stay in operation until replacement, the possibility of failure of those units has to be included. Then, the determination of system state probabilities has to be done iteratively.Practical implicationsThe methodology proposed in this paper enables comparison of different designs for a float system submitted to periodic overhauls through some different performance measures. System state probabilities can also be used to determine the idle time of units, which is essential to obtain the cost of production loss and, therefore, the system maintenance cost.Originality/valueThis paper considers simultaneously three different factors that influence maintenance float system performance: the number of spare units and the number of maintenance crews (which mean minimising the downtime effect) and the overhaul frequency (which means decreasing the number of failures).