Construction encompasses a wide range of projects, from small house renovations to significant building projects. The major investment in Ethiopia's construction industry is in the transport infrastructure. Construction investments affected by time and cost overruns. To effectively finish a project on schedule, it is crucial to be able to anticipate time properly. As "time is money," a variety of developing duration estimation technologies has been created to produce precise predictions. The main objective of this study was to develop a mathematical model that can be used to predict the construction duration of highway projects with accuracy. A literature review and an interview were conducted in this study to investigate appropriate factors that affected the construction duration prediction mathematical models of highway projects and the variables were collected through a desk study. The factors of construction duration prediction mathematical models of highway projects were the project's cost, project location, weather conditions, tender type, pavement type, project financier, road length and road width. The factors related to thirty highway projects used to develop the models were constructed in Amhara Region, Ethiopia from 2002–2013 E.C. The data sets used for the study were obtained from the Ethiopian Road Authority and Amhara Road Construction Enterprise. For estimating construction durations, models have been developed to improve upon this approach; all of them were conducted. The estimation accuracy and correlation of the regression analysis were investigated by using coefficient of determination, mean absolute percentage error, mean squared error and root mean squared error. Based on this value, the artificial neuron network has more accuracy and goodness of fit than other prediction mathematical models in this study. Hence, the artificial neuron network model was found to be the best alternative. Finally, compare the study with previous studies. Recommend to the transportation sector to use artificial neural network specially the road sector. Keywords: construction, Duration, Prediction, Models, Highway, Amhara