Correlation Filter-based Trackers have demonstrated impressive performance in object tracking, surpassing traditional methods in various benchmarks. However, the tracking process undergoing deformation, rapid motion, or occlusion remains a significant challenge. The repetitive nature of training sets in correlation-based filter tracking often leads to undesirable effects at boundaries, hindering tracking efficiency. To address these limitations, this paper proposes a novel hybrid attention-based correlation filter method (Att-DCF). By employing a discrimination filter and mitigating the impact of irrelevant features, our approach achieves accurate and stable target localization. Att-DCF is evaluated on the recent standard three common benchmarks: OTB-100, Temple-Colour128, and UAV123. Results indicate a notable achievement of 2.71 \(\%\) in area under the curve and 1.38 \(\%\) in precision, on OTB-100. Similarly, on Temple-Colour128 and UAV123, Att-DCF outperforms the baseline DCF tracker by 3.99 \(\%\) and 3.43 \(\%\) , and 4.72 \(\%\) and 1.76 \(\%\) , respectively.