A new control approach is proposed for the probability density function (PDF) control of non-Gaussian stochastic systems, using PI controllers. Using the square root output PDF model and the weight dynamics. the PDF tracking is transformed to a con-strained dynamical tracking control problem for weight dynamics, where LMI techniques are used to design a generalized PI controller such that stability state constraints and tracking performances can be guaranteed simultaneously. (C) 2004 Elsevier Ltd. All rights reserved.