AbstractEnvironmental DNA (eDNA) measurements have been widely applied in the biomonitoring of organisms. However, different DNA extraction methods may cause changes in yield and consistency, resulting in an inaccurate estimation of eDNA, especially when quantitative measurements are performed. This study focused on evaluating the effects of DNA extraction methods on the magnitude and consistency of fish DNA yield from water samples. Samples were collected periodically over a year from river and lake water systems, and reference eDNA was spiked into them. The spiked samples were extracted using three methods: using lysis buffer for blood and cells (AL method) with the DNeasy Blood and Tissue Kit (Qiagen), using lysis buffer for tissue (ATL method), and the microfluidic‐channel method (BC method). DNA yields were compared by determining the copy number using qPCR. The ATL method showed better stability and a higher yield than the AL method. In addition, despite being comparatively simple, the BC method performed extractions stably and at relatively high yields. Based on the long‐term evaluation of eDNA measurements using various methods, we showed that differences in DNA extraction methods may alter DNA yield and stability.