This study proposes an effective thermal control for plastic injection molding (polymer: Santoprene 8211-45 with density of790kg/m3, injection pressure: 1400psi (9,652,660 Pa)) in a laminated die. For this purpose, a comprehensive control strategy is provided to cover various themes. First, a new method for determining the optimal sensor locations as a prerequisite step for modeling and controller design is introduced. Second, system identification through offline and online training with finite element analysis and neural network techniques are used to develop an accurate model by incorporating uncertain dynamics o f the laminated die. Third, an additive feedforward control by adding direct adaptive inverse control to self-adaptive PID is developed for temperature control of cav ity wall (cavity size: 52.9 x 32.07 x 16.03 mm). A verification of designed controller's performance demonstrates that the proposed strategy provides accurate online tempera ture tracking and faster response under thermal dynamics with various cycle-times in the injection mold process.