In this paper, we introduce the concept of extended linear negative quadrant dependence (ELNQD, in short). We establish a new concentration inequalities and complete convergence for the distribution of sums of extended linear negative quadrant dependent random variables. Using these inequalities for proved the complete convergence of first autoregressive processes model generated by identically distributed ELNQD errors.