Investigating natural processes in arsenic (As) polluted areas and plants that have naturally chosen to grow there pose practical restoration recommendations. This study aimed to assess long-term changes in natural As dendroremediation dynamics for Quercus robur L. and Tilia cordata Mill., tree species capable of growing in areas polluted by mining activities. We examined total As and its forms, as well as B, Ca, K, Mg, Na and P, in soil and trees over 5 years. We also characterized pH and EC of soil, examined proline content in tree organs, and calculated Bioconcentration Factor (BCF) and Translocation Factor (TF) for As.Initial As concentrations in soil were 37.0 mg kg⁻1 under Q. robur and 34.7 mg kg⁻1 under T. cordata, significantly decreasing after 5 years to 10.5 mg kg⁻1 and 9.51 mg kg⁻1, respectively. This corresponds to pollution reduction of up to 71.8%. A notable decrease in As(III) and dimethylarsinic acid, along with increase in other organic As forms in soil, was observed. Additionally, concentrations of essential elements in soil, as well as its pH and EC, decreased over time.Both tree species accumulated substantial amounts of As in their organs, but the dynamics of this process were species-specific. During first 4 years, T. cordata accumulated more As and exhibited higher BCF, but in the 5th year, it was clearly surpassed by Q. robur. The highest TF was calculated for Q. robur in year 3, and for T. cordata in years 2 and 3. Generally, limited aboveground movement of As was indicated: BCF > 1 were calculated for years 2 and 3, while TF were consistently < 1. Proline content increased significantly in all organs, correlating with As, especially in Q. robur. In contrast, Q. robur leaves mapping revealed stable macroelement distributions, but clear variations were observed for T. cordata., which may suggest specific reaction to stress.These findings suggest that both species can effectively restore As-polluted areas, though with different dynamics. The selection of species for dendrorestoration should be based on whether the goal is faster remediation with lesser overall reduction (e.g. T. cordata) or slower remediation with ultimately greater pollution reduction (e.g. Q. robur).