A method for establishing a dynamic maintenance scheduling tool for a specific part of a machinery based on condensed prior knowledge of the part of the machinery in a population of machineries. The scheduling tool is in turn used in a method for establishing a dynamic maintenance schedule for a specific part of a specific machinery, wherein parameters related to the usage, including relevant parameters representing factors influencing the lifetime of the specific parts, are utilized as input data to the dynamic maintenance scheduling tool for the specific part of the machinery, whereupon a dynamic maintenance schedule for the specific part of the specific machinery is achieved as output data from the scheduling tool. The method is in particular dedicated to industrial robot systems.