Rainfall data is essential for any study related to water resources. Daily rainfall has its specific characteristics which is continuous time series and discrete data with non-normal distribution. Generally, methods to estimate missing daily rainfall data, for examples arithmetic mean, inverse distance weighting method (IDW) still have some limitations. Such founded limitations are: 1) underestimate of average daily rainfall, 2) overestimate of non-zero rainfall events, and 3) underestimate of extreme rainfall magnitude. This study attempts to develop an imputation method for daily rainfall using quantile approach (QT) which is based on Bernoulli-Gamma distribution, and then compare to IDW method. The study results reveal that QT could yield sample statistics such as maximum, mean, and variance of estimated daily rainfall better than IDW. In addition, the 95th and 99th percentiles of rainfall depths from QT method are closer to the observed data. Therefore, QT method is capable to estimate extreme rainfall magnitude superior than IDW approach. Moreover, QT gives a higher accuracy in number of zero and non-zero rainfall events. Using QT method might be appropriate for any study that concerns with extreme rainfall events since QT would give more accurate results.