Small fragments of artificial polymers (microplastics, MPs) has been reported for multiple environmental matrices from our planet. The omnipresent existence of these microplastics even in remote polar areas have raised concern about their potential environmental impacts and created a need for effective and standardized analytical methods targeting their detection in environmental samples. So far, no methods have been developed for detecting microplastics in organic-rich soils. In this master thesis, I evaluate two analytical methods (both based on canola oil extractions) targeting microplastics in two contrasting soil matrices; one mineral rich (sandy mineral soil from a Podzol) and the other by organic matter (sample from a Histosol). I hypothesize that the detection of microplastic has a bias that depends on specific plastic particle properties (size, polymer type and morphology) as well as on the organic content of soil samples. My results show that the recovery of added plastics is strongly dependent on particle size and diminishes with decreasing microplastics length. This result was repeated by both extraction approaches.  Polymer shape and soil characteristics (organic matter content) affect MPs recovery if oil extractions are conducted without pre-treatment (oxidation) step. Here, fibers proved most difficult to detect and low recoveries suggested that the method was not applicable to organic rich samples. The addition of a pretreatment step including oxidation with sodium hypochlorite improved recoveries for organic rich samples and removed the effect of soil type and polymer shape. Hence, the use of a pretreatment is essential to extract MPs from organic-rich soils, but it also decreases the overall recovery for all type of studied polymers and mostly fibers. My study suggests that there is a substantial bias when detecting MPs in soils that is causing a general underestimation, especially for small, fibrous particles in organic rich soils.