Abstarct: Positive quadratic dependence (PQD) is related to the joint behavior of random variables; when small (or large) values are likely to be observed for them simultaneously. Recognition of the presence of this dependency structure in the data can lead to improved computations and decisions. Therefore, in this thesis, the independence test of two random variables versud having PQD dependence has been studied. Given that the actual distribution of random variables is practically unknown, we will use their empirical versions in a nonparametric approach baxsed on the copula function. The simulation results show that the proposed empirical likelihood test performs well, having the ability to reveal (X,Y) are PQD.