AbstractTraffic flow is a continuous phenomenon. The irregular patterns in the traffic flow data show the complexity of the system under a variety of internal and external factors of restrict and influence. As such, local mean prediction method (LMPM) has been employed in order to do prediction. Firstly, traffic flow data has been reconstructed to m-dimensional phase space. Next, the phase space is used to predict traffic flow using chaos approach. The traffic flow data reveals chaotic behaviour through analysis using Cao method. The prediction performance shows that the value of correlation coefficient (CC) is greater than 0.5 for all stations that have been analyzed. The value of CC is good and acceptable due to the complexity of the system. These findings are expected to assist in fulfilling the Road Safety Plan 2014-2020 by the Ministry of Transportation as well as local authorities in traffic management...