Phytoremediation of petroleum hydrocarbons in subarctic regions relies on the successful establishment of plants that stimulate petroleum-degrading microorganisms, which can be challenging due to the extreme climate, limited nutrients, and difficulties in maintaining sites in remote locations. A long-term phytoremediation experiment was initiated in Alaska in 1995 with the introduction of grasses and/or fertilizer to petroleum hydrocarbon (PHC)-contaminated soils that were subsequently left unmanaged. In 2011, the PHC concentrations were below detection limits in all soils tested and the originally planted grasses had been replaced by volunteer plant species that had colonized the site. Here, we sought to understand how the original treatments influenced the structure of prokaryotic communities associated with plant species that colonized the soils and to assess the interactions between the rhizospheric and endophytic communities of the colonizing vegetation 20 years after the experiment was established. Metataxonomic analysis performed using 16S rRNA gene sequencing revealed that the original type of contaminated soil and phytoremediation strategy influenced the structure of both rhizospheric and endophytic communities of colonizing plants, even 20 years after the treatments were applied and following the disappearance of the originally planted grasses. Our findings demonstrate that the choice of initial phytoremediation strategy drove the succession of microorganisms associated with the colonizing vegetation. The outcome of this study provides new insight into the establishment of plant-associated microbial communities during secondary succession of subarctic areas previously contaminated by PHCs and indicates that the strategies for restoring these ecosystems influence the plant-associated microbiota in the long term. IMPORTANCE Subarctic ecosystems provide key services to local communities, yet they are threatened by pollution caused by spills and disposal of petroleum waste. Finding solutions for the remediation and restoration of subarctic soils is valuable for reasons related to human and ecosystem health, as well as environmental justice. This study provides novel insight into the long-term succession of soil and plant-associated microbiota in subarctic soils that had been historically contaminated with different sources of PHCs and subjected to distinct phytoremediation strategies. We provide evidence that even after the successful removal of PHCs and the occurrence of secondary succession, the fingerprint of the original source of contamination and the initial choice of remediation strategy can be detected as a microbial legacy in the rhizosphere, roots, and shoots of volunteer vegetation even 2 decades after the contamination had occurred. Such information needs to be borne in mind when designing and applying restoration approaches for PHC-contaminated soils in subarctic ecosystems.