Construction time prediction is a very crucial part of project planning, as several key decisions are based on the estimated durations. Funding, financing, and resource allocation decisions are based on the early estimated durations. This research was aimed to develop a construction time prediction model for public building projects in Addis Ababa, Ethiopia. The study incorporated a total sample of fifty-six public building projects completed from 2007 to 2014 G.C. The result revealed that Bromilow’s Time - Cost (BTC) principle was moderately applicable for the sample data set. However, the cubic regression model (CUB) was found to provide a better time-cost relationship. In contrast, the study has shown a poor relationship between actual time and gross floor area. Furthermore, stepwise multiple linear regression (MLR) analysis consists of three statistically significant variables; Average floor area (Av.A), Height of building (H), and Cost level (CL) were found a better fit time model. The final selected time model (i.e. MLR) was validated for regression assumptions, Mean Absolute Percentage Error (MAPE), and Sum of Squared Errors (SSE).

Keywords: Bromilow time-cost principle, Time prediction model, Regression analysis