This paper analyses the optimal nominal tuning of a new modification of predictive disturbance observer (PDO) based filtered PI control (PDO FPI) applied to a first order plus dead time (FOPDT) plant with exactly known parameters. The impacts of applied filters on optimal controller tuning and on achievable closed loop performance are evaluated first of all. The limits of achievable performance are compared with those of traditional two degree of freedom (2DOF) PI control, with both controllers tuned by the multiple real dominant pole method. This comparison shows the potential of PDO FPI control to improve tracking and regulatory dynamics significantly, permitting the Pareto-like servo/regulator trade-off of 2DOF PI control to be removed. Two PDO FPI tuning approaches are proposed, allowing optimal filter degree and time constants to be evaluated. The first tuning scenario considers optimization of closed loop performance expressed in terms of the Integral of the Absolute Error (IAE) weighted alternatively by the relative total variance TV1 of the control signal. This is carried out by changing the filter order n under the constraint that a constant position of the dominant closed loop pole is maintained. This keeps the dynamics of the setpoint step responses almost unchanged. In the second tuning scenario the optimization is carried out under a constraint on constant speed of disturbance step responses. All the main results are then numerically checked for the integral first order plant with dead time by the performance portrait (PP) method. The analysis presented here shows that the new PDO FPI structure substantially enriches the spectrum of controllers applicable to simple control tasks. (C) 2013 Elsevier Ltd. All rights reserved.