The objective of this paper was to develop a new method to predict topsoil (0-40 mm) porosity (TSP) using soil surface roughness (SSR) under rainfall influence. For this purpose, serial experiments were conducted in Dikopshof Experimental Station, Bonn University, Germany at each summer form year 2006 to 2009. In these experiments, TSP and SSR of two fields cultivated by two different tillage types (moldboard plow + harrow for field 1 and rotary hoe + harrow for field 2) were measured after tilled (0-50 days) under natural conditions. In here TSP was measured by pycnometer and SSR by laser profilemeter. A new index of rainfall was defined as accumulative mean rainfall (ARF) to investigate the rainfall influence. The four year’s results showed that TSP reduced with the increase of ARF and decrease of SSR. Based on this relation, a multiple linear regression model for predicting TSP using SSR under rainfall impacts was established. Verifying by Ftest, this model was suitable for estimating TSP. Using the data of laser profilemeter and rainfall, this model will be used to predict topsoil porosity under different tillage types.