Public works are constructed for use by the public. In a construction project's preliminary design phase, the pressure of ensuring the quality and timely completion of the construction project often compels the contractor, who may not possess much related information, to estimate the schedule according to empirical examples of similar projects. This reason alone accounts for a considerable proportion of construction delays in Taiwan. Although public agencies store a tremendous amount of information arising from public tenders each year in the public tender management systems currently operational in Taiwan, few attempts have been made to identify any discernable patterns through data analytic techniques. In consideration of this, the present study seeks to establish a model integrating tender histories from the Directorate General of Highways for scheduling construction projects still in their preliminary design phases. The model will also be of reference to the launch of early warning systems for construction scheduling in Taiwan. The present study analyzes, through data mining techniques, data on the 2,413 public tenders, public tender construction audit reports, and 90 road construction tenders stored in the database of the Directorate General of Highways' closed-contract public tender management system between 2008 and 2012. The results of this analysis are used to propose a model for scheduling construction projects still in their preliminary design phases. The systematical approach is expected to facilitate expedited schedule estimation and verification.