This paper investigates a defect appearing in 鈥淥bserver-based finite-time control of time-delayed jump systems鈥? which the observer-based finite-time controller via state feedback could not ensure stochastic finite-time boundedness and stochastic finite-time stabilization with the prescribed control index of the resulting closed-loop Markov jump systems. The corrected results are given and the improved optimal algorithms and new simulation results are also presented in this paper.