In Europe, muskrat is an invasive species that can profoundly affect lake ecosystems. We developed a landscape-based prediction model for the occurrence of muskrat based on 237 muskrat and 236 randomly selected lakes within the distribution range of muskrat in northern Sweden. We analyzed the importance of slope and cover of vegetation types for the occurrence of the muskrat at 12 spatial scales (50–1000 m) from the lakeshores. Logistic regression models that incorporated slope and percentage cover of swamps, meadows and lakeshore meadows successfully predicted the occurrence of muskrat. The importance of the predictor variables changed with increased distance from the shoreline. Our results were confirmed with an independent data set (n = 29) from the southern distribution range of the muskrat in Sweden. The prediction model can be used to assess the risk of muskrat occurrence in lakes as well as for the development of muskrat-related conservation measures.