Abstract: In this work a special attention is devoted to the robustness study of a biased estimator RB(β,γ) having two easily tuned parameters. In the chemical and biochemical reaction engineering systems, which undergo inevitably the random disturbances, the proposed estimator, as compared with the least-squares estimator and ridge estimator, is significantly better in both estimate accuracy and convergence speed in the case of dealing with the problem of nearly-singular design matrix due to its robustness.