One of the estimating of density function which has been recognized is histogram. Histogram has some weaknesses, i.e. the different starting points and the width of class intervals. Different starting points or different class intervals result different histogram forms. This article is about the estimating of the density function by using kernel function. This method does not require the determination starting points and the interval class width. The obtained curve has a smooth density function, a small sampling variance, and the important information from data are still kept.