The funds allocated for the maintenance and repair of the public buildings must be spared as scheduled, because of the limited financial resources. However, because of the uncertainties embodied within the very nature of the maintenance and repair works, and the working load, most of the time, the funds based on the untruth-worthy bill of quantities remain below the approximate cost and this causes delays in the works or a complete collection of projects and returning the allocated funds back to the budget eventually. On the other hand, due to discounts, the tender prices usually remain within the limits of the funds allocated at the very beginning. In this paper, for the purpose of using the limited funds allocated for the maintenance and repair works effectively, by taking the factors effecting the bidding into consideration such as; approximate cost, and the city where the work is to be done, and contracting authority, and repair type, and bidding date, and duration of the work, and number of accepted bids, the correct estimates of contract prices have been studied. In this regard, the 211 bidding data concerning various maintenance and repair projects carried out by a public institute in 2015 have been analyzed using regression-correlation method. The results of the analysis have been tested by way of Artificial Neural Networks (ANN) and some comparisons made. It has been shown that the results of the study will lead the way for public officials in utilizing the funds allocated for the maintenance and repair of public buildings more effectively and efficiently.