This paper presents a novel method for pole assignment with robustness measured in terms of the spectral condition number of the closed-loop eigenvector matrix. It is established that the spectral condition number can be minimized asymptotically via a sequence of unconstrained minimizations on some auxiliary objective functions. Moreover, the sequence of minimizers converges to a minimizer of the spectral condition number. A numerical algorithm with analytical formulas of the gradient of the auxiliary objective functions is provided. The efficiency and effectiveness of the approach is demonstrated via an example.