The thesis is concerened with a bus fleet maintenance study for an intercity bus company in a deeloping country, namely, Malaysia. Ideally, operators wuld wish to have their buses available to operate in a good condition 100 percent of the time. Often, howeer, buses do fail whilst thery are in service which can entail both cost and downtime as well as passengers' dissatisfaction. It is of prime concern to manu operators (as to the bus company) that, in order to remain competitive, such failure occurences must be reduced while controlling the total cost and downtime. In addressing such operator's concerens and, in general, the maintenance decision problems within the context of the bus company, this study applied the modelling approaches of snap-shot modelling for problem identification and definition and delay time modelling for evaluating the effectiveness and efficiency of the existing and alternative maintenance practice. It is found that, in a data starved situation sucas that considered here and believed to be typical of developing countries, both the snap-shot and athe dealay time modelling are particularly appropriate and applicable. The use of snap-shot modelling has enable the management of the company to see, for the first time, a comprehensive picature of their bus operation and behaviour, which resulted in several actions which led to improvements. The use of delay time modelling technique has enabled issues such as whether or not the exixting level of light inspection is appropriate, whether or not twolight inspection schedules should be adopted in place of the existing single light inspection schedule, whetjer or not a different policy should be applied for a different bus sub-fleet, and whether or not notable benefit would follow an improvement of the existing inspection pracatice, to be addressed and evaluated.