The study is concerned with the nonparametric kernel estimation to determine the soil porosity and pore size distribution. The kernel density estimation, the kernel estimation of cumulative distribution function, and the kernel estimator of quantile are considered. After a short description of the method, practical aspects and applications in agricultural science are presented. The nonparametric kernel estimation does not require a priori assumptions relating to the choice of the density function shape. Moreover, its natural interpretation together with its suitable properties makes them an adequate tool among others in estimation methods.