Pole-placement is one of the oldest control design method used to modify the behavior of a given plant dynamics. In the literature, analytical solution of output feedback pole placement is available only for two-input two-output systems. For general linear time-invariant, multi-input, multi-output (MIMO) systems, static (constant gain) output feedback pole placement is an open non-deterministic polynomial-time (NP) hard problem. This paper proposes a semi-analytical solution of constrained (minimum number of nonzero gains) output feedback pole placement problem in MIMO systems. To get the semi-analytical solution, first an analytical formula is developed for pole-placement in single-input, multi-output (SIMO) systems. This formula is an extension of Bass-Gura's formula used for state feedback pole placement in single-input systems. Along with the gains needed for partial pole-placement, this novel formula computes the coefficients of characteristic polynomial corresponding to the remaining poles of the plant. Using this analytical formula a novel iterative procedure is developed to place all the poles of a multi-input system using output feedback. Proposed iterative method needs only the solution of linear equations at each step of iteration. Starting from the known initial guess, it searches for the solution in a lower dimension space compared to other numerical methods. The algorithm computes a full rank feedback gain matrix. For a decoupled m-input system the proposed method computes the output feedback gains in exactly m-steps, non-iteratively.