Construction projects’ cost and degree of complexity are mounted rapidly in North Africa and Arabian Gulf Countries inducing higher potential levels of risks to costs and time overruns; therefore, greater attention must be directed to control both time and cost. The conventional techniques for estimating activity’s duration and time scheduling have been proved themselves as unreliable techniques in front of growing construction projects’ challenge. Simulation techniques have been utilized to provide reliable solutions for risky construction project scheduling. Beta and Normal Probability Density Functions (PDF) have been widely utilized in the past to represent the activities’ duration included in simulation process. This paper presents a statistical analysis for construction activities estimated and actual durations in North African and Arabian Gulf countries. The analysis was carried out on twenty construction projects including more than 125,000 activities representing all engineering, procurement and construction (EPC) activities. The statistical analysis of results proved that the most appropriate PDF to represent the duration of activities is the Lognormal distribution followed by the Gamma distribution. However, utilizing Beta or Normal PDF may be utilized at lower confidence level. The analysis manifests that about 40% of the activities contribute to the delay of the whole project.