Development of a civil aircraft maintenance program during early design stage has been carried out by Maintenance Steering Group, which is a team of manufacturer, authority and customers, according to a specified methodology which requires very high experience compare to other design steps. Main sources of maintenance design process are based on pilot reports, maintenance records, failure alerts and manufacturers' recommendations and requirements. During this process, one of the most important topic is task related items such as task types, intervals and durations. Maintenance task durations or repair time are very important for airline companies because availability of an aircraft directly related with this parameter. Aircraft maintainability allocation which is a process to identify the allowable maximum task time for each aircraft component or system is based on mostly experience and out of design office's control. In this study, a new method with two steps has been developed to create an alternative technic for experimental ones. At the first step an existing methodology developed for maintenance allocation has been improved by using a different technic. Improved method shows that newly established correlation between aircraft systems and task times has very high coefficient of determination compare to the existing method. At the second phase of the study several quantitative analysis have been performed by examining 1175 maintenance tasks which are accepted as standard maintenance actions by aviation industry, coming from Maintenance Steering Group methodology and six weight factors have been established for the new method. By using feed forward artificial neural networks for newly identified weight factors, maintenance task allocations has been established. Results shows that newly proposed method can be applicable for any maintenance process during early design stage. However since this study focused on system and component tasks in Maintenance Steering Group, a different perspective is required for structural and zonal tasks allocations.