Three-tank benchmark system is a nonlinear multivariable process that is a good prototype of chemical industrial processes. The open loop process is unstable inherently, thus a proper controller must be applied to stabilize the system as well as reaching good dynamic response. In this paper, several controllers have been proposed using pole-placement with integral control. Then, the best controller is selected considering deterministic and stochastic performances. Deterministic performance was evaluated using step test on closed loop. Minimum variance (MV) index was used for stochastic performance assessment. Multivariate time series (MTS) is cached from sampled outputs of the simulated system in closed loop. Then, maxlikelihood method is used to fit an ARMAV model to the MTS. The obtained model is used for determining MV index. This approach can be used for determining the efficiency of loop after implementation.