In the recent century, a process of urbanization has increased globally. Previously rural or natural land have been converted into impervious surfaces to make way for housing, industries, and roads. This anthropogenic process has resulted in an increased amount of surface runoff from precipitation, so-called stormwater. Stormwater can accumulate a range of pollutants when it flows over the impervious surfaces of our cities. These pollutants can have a negative impact on the lakes and streams that receive the stormwater from the urban environments. To reduce the environmental problems associated with the content of stormwater, various techniques have been developed for stormwater treatment, with the aim of reducing the pollutant load in the runoff before it enters the receiving waterbody. One such technology is stormwater biofiltration, also known as bioretention. Stormwater biofilters were developed in the USA in the early 90's and they utilize the natural water remediation properties of plant-soil systems. They are generally characterized by a vegetated submerged filter bed with an underlying drainage layer. They have shown to be an effective method for stormwater pollutant removal. The treatment processes take place both in the vegetation and in filter material. As th ematerial choices and design of the biofilters can vary, so can its treatment performance. Stormwater biofilters have grown in popularity in the last decades since their development and numerous studies have been conducted to evaluate the systems’ treatment efficiency. However, knowledge gaps still exist regarding their implementation in colder climates and the suitability of different configurations and materials. This study examines the removal performance of total and dissolved heavy metals (Cd (cadmium),Cu (copper), Pb (lead) and Zn (zinc)), phosphorus, nitrogen and total suspended solids (TSS) in three stormwater biofilters in a Swedish climate, located in central Malmö. The current biofilters are designed with different configurations of their filter media and are built with 1) sand-basedfilter material 2) sand-based filter material with a submerged zone and 3) filter media consisting of 50% sand-based material in combination with 50% pumice. The results showed that the treatment capacity of the biofilters with a filter media of only sand (biofilter S) and with sand as well as a submerged zone (biofilter S_SZ) was similar for all pollutants. The reduction of total levels of metals (> 85 %) and TSS (>90 %) was consistently high and similar to levels achieved in previous studies for both temperate and colder climates. The removal of dissolved metals was lower in comparison to the removal of the total metal fractions, but the dissolved fractions were still generally reduced in the effluent. A positive removal of total phosphorus and total nitrogen was overall displayed in the effluent from the two biofilters; however,leaching was shown for the dissolved fractions. For nitrogen (N) species, the concentrations in the runoff were generally below the detection limit for the analysis making it difficult to establish probable removal percentages. For the biofilter S_P, which contained a mix of sand and pumice, the removal capacity was lowerfor all parameters compared to the other designs. Overall, the pollutant removal performances are regarded to be satisfactory for biofilters S andS_SZ and their implementation suitable for the given site. However, further investigations should be performed during warmer seasons, especially regarding the removal of nutrients.