Abstract: This paper presents two versions of state-space self-tuning controller for multivariable stochastic systems having the same number of input and output, namely, the multivariable statespace pole-placement self-tuning controller and the multivariable state-space tracking self- tuning controller. The paper gives not only the construction of both controllers, but also the theoretical analysis of their algorithms. The former algorithm can be applied to stable/unstable, minimum/non-minimum phase linear time-invariant multivariable system and the latter can he applied to stable/unstable, minimum phase linear time-invariant multivariable systems. The simplicity of construction and better control performance of the proposed algorithms facilitate their practical application to real systems.