The paper introduces the R package `relaimpo`, which implements six methods for assessing the relative importance of regressors in linear regression models. Among these, the averaging over orderings (Lindeman, Merenda, and Gold, 1980) and the newly proposed metric (Feldman, 2005) called `pmvd` are highlighted as the most computer-intensive and recommended methods. The package also provides bootstrap confidence intervals for these metrics. The paper uses the `swiss` dataset to illustrate the methods and functionality of `relaimpo`. It covers simple relative importance metrics such as `first` and `last`, as well as more complex metrics like `betasq` and `pratt`. The paper discusses the limitations of these metrics and explains how `relaimpo` addresses them. Additionally, it covers the calculation of relative importance based on the covariance matrix and bootstrapping regression models. The paper compares `relaimpo` with other R packages and highlights its advantages in terms of computation time and confidence interval computation.The paper introduces the R package `relaimpo`, which implements six methods for assessing the relative importance of regressors in linear regression models. Among these, the averaging over orderings (Lindeman, Merenda, and Gold, 1980) and the newly proposed metric (Feldman, 2005) called `pmvd` are highlighted as the most computer-intensive and recommended methods. The package also provides bootstrap confidence intervals for these metrics. The paper uses the `swiss` dataset to illustrate the methods and functionality of `relaimpo`. It covers simple relative importance metrics such as `first` and `last`, as well as more complex metrics like `betasq` and `pratt`. The paper discusses the limitations of these metrics and explains how `relaimpo` addresses them. Additionally, it covers the calculation of relative importance based on the covariance matrix and bootstrapping regression models. The paper compares `relaimpo` with other R packages and highlights its advantages in terms of computation time and confidence interval computation.