This paper is concerned with the optimal control problem for continuous-time systems with general noises, based upon sampled data, under the quadratic cost functional. The system is described by a linear stochastic integral equation, the observations are made at discrete times, and the noise processes are not assumed zero-moan Gaussian and/or white. Derived is the optimal controller algorithm, where the optimal input consists of the following two: (1) the optimal input for usual linear-quadratic-Gaussian control systems and (2) its correction input due to the fact that the noise processes arc non-white and have non-zero means.