Abstract Management of maintenance floats is an important aspect of decision-making in most service and manufacturing organizations. Such decision involves the procurement or leasing of standby units, and the maintenance of repair shops in order to satisfy the organizations's demand load. It also interfaces with capital budgeting decisions since an optimum combination of standby units and repair personnel that satisfy both the organization's target service level and budgetary conditions must be determined. This paper proposes a framework to analyze maintenance float problems through the use of regression metamodels in simulation. We consider a maintenance float problem with exponential services and study it using a 2 4-1 IV fractional factorial design. The model generated is validated and shown to be quite effective. Application in decision analysis is illustrated. The regression metamodeling approach presented here is also applicable to a wide variety of complex problems where simple analytical solutions may not be easily obtained.