In this work, the Adsorption Method for Sampling (AMESA) is developed to collect long-term samples of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) from stacks. However, the variation of in duration might affect the analytical results due to the memory effect in the AMESA. This research is one of the very few reports of using AMESA in sinter plants. Eight long-term AMESA (LA) samples were collected, followed by 16 parallel short-term AMESA (SA) and manual (SM) samples. The LA reports good correlations (R = 0.82–0.83) between PCDD/F emissions and the recovery feedstock, which can be further used for emission reduction. However, the breakthrough effects of the highly chlorinated PCDD/Fs is in the range of 6–14% in LA samples, and these are further improved by increasing the glass wool and adsorbent (with higher specific surface area, lower pore size, and higher polarity) packing, and finally reduced to 5.17%. For the short-term sampling, a maximum difference of 170% occurred between SA and SM, if the sampling strategy suddenly changed from LA to SA. The PCDD/F memory effect remained for the first 8 SA samples, with the dominant 1,2,3,4,6,7,8-HpCDD and OCDD. Therefore, five post-washings for the AMESA system took place to successfully reduce the PCDD/F residual mass to > 0.2% of a normal LA. Unfortunately, these residuals were still equal to 51.4% of a normal SA. Therefore, it is not suggested that AMESA replace the current manual sampling method without any modification. According to this study, both the breakthrough and memory effects could be inhibited. Finally, a new inspection system, including both AMESA and short-term samples, is then proposed for improving the PCDD/F control strategy in Taiwan.