Every natural behaviour is non-linear, but not always is chaotic. This paper aims to investigate low dimensional chaotic behaviour of study area: Konya Basin by using non-linear time series techniques with three stages: i) Mutual Information, ii) False Nearest Neighbour ( FNN) algorithm, iii) Stretching Exponential. These techniques calculate the delay time, the embedding dimensions and the maximal positive Lyapunov exponent respectively. The data set consists of daily average flow rates of three stations in Konya Basin through the study period between 1968 and 2014. Analysed data implied that these time series have shown chaos. This information helps catchment manager to forecast future and the extreme flow rates such as droughts and floods.