In thermoplastic tape-laying, a laminated composite is constructed by local application of heat and pressure. Each new layer is bonded to the previously bonded layers (substrate). The temperature at the interface between the top ply and substrate is critical to achieving interlaminar bonding. Recent research on the in-situ thermoplastic composite tape-laying process has focused on the modeling, numerical analysis and experimental analysis, but little research has considered the control of this process. In this work, a methodology is proposed for modeling and control of in-situ thermoplastic composite tape-laying. The key to the control algorithm is predicting the temperature at the interface between top ply and substrate. Based on a process model, a state feedback controller and a state estimator for temperature are designed for closed-loop control using the linear quadratic method. Two different approaches are used to develop the process model for real-time closed-loop control through temperature feedback. In the first approach, a low order lumped model is constructed from a finite difference scheme. The second approach constructs an empirical model through system identification. The structure of the two models are identical, but the parameters differ. The experimental results have shown a significant improvement in temperature control compared with the previous control method indicating that the proposed modeling and control methodology can produce a high quality thermoplastic composite laminate.