This paper reports on an investigation of a highly-accurate method for estimating the one-minute rain rate distribution for various locations throughout Japan using observed precipitation data collected through the Automated Meteorological Data Acquisition System (AMeDAS). In recent years, there has been an increasing tendency for very hard rain to fall in a short period of time, and the relationship between one-hour and one-minute rain rate distribution has changed in comparison to the one-hour and one-minute precipitation data collected during the past 25 years. As an example of the reflection of this tendency on the highly-accurate method for estimating the one-minute rain rate distribution, we present a method that estimates the one-minute rain rate distribution by supplementing the hourly rainfall data with 10-minute rainfall data and show quantitatively an improvement in the accuracy over existing estimation methods.