Residual stresses are originated during manufacturing processes of metallic materials, so its study is important to avoid catastrophic accidents during component service. There are two main types of residual stresses, according to the length scale; macroscopic and microscopic. While the determination of tmacroscopic ones is almost a routine analysis, determining the microscopic stress of individual grains remains a pending task. In this paper, we present an approach using genetic programming to obtain the micro residual stresses in grains of a quenched cylindrical sample of aluminium alloy AA5083. The microstructure of this alloy is formed by grains with different orientation and stress. To obtain the stress of each grain we estimate the values of the micro residual stresses for each crystallographic orientation using information from neutron and electron back-scattered diffraction experiments. This information includes orientation maps of a normal section to the cylinder axes (individual orientations) and the dimensions of each grain. We assume that the micro residual stresses of each grain can be expressed as a function based on these variables and use genetic programming to find this expression.