This paper considers the problem of adaptive LQG control design when the system control input is subject to an amplitude constraint. A first-order stochastic system is considered mainly because its output can be satisfactorily described by a Gaussian probability density function (PDF) or can be calculated numerically. In this light, two algorithms are proposed. In the first algorithm an approach given by Toivonen has been adopted where the PDF of the output is approximated by the Gaussian density in order to develop a suboptimal adaptive feedback controller. The approach in the second algorithm is based on the numerically calculated PDF of the system output yielding a near-optimal feedback gain. The proposed algorithms are compared with existing ones via simulations.