This work contributes on how a parameter optimization scheme tackles the system identification problem in type 1 diabetes (T1D) patients to derive a dynamical model with potential application on feedback control schemes for blood glucose regulation. That is, the contribution aim is to identify values of system (sensitive) parameters such that a patient-specific model is derived towards the future design of feedback control for assisting T1D therapy. To this end, a differential equation system is proposed to model the blood glucose dynamics in T1D. A challenge in control systems regards the system identification to capture suitable response from available measurements of the blood glucose levels. T1D is particularly interesting due to the glycemia inter-variability. This fact has been recently highlighted because of the continuous glucose monitoring has revealed the need of specificity at glucose dynamics model for every single patient. Hence, a class of artificial intelligence algorithms are performed towards the identification of a control system for the individual glucose metabolism. Here, three AI algorithms perform the system identification towards future implementation. Each one of the three AI algorithms comprises two parts: a physiological model and a parameter optimization scheme to capture the time response from a set of glucose data obtained from measurements. The explored AI algorithms for the parameter optimization are respectively approached via Genetic Algorithm, Particle swarm optimization algorithm, and Taguchi sliding based differential evolution algorithm. A set of patients allows us to explore experimentally the performance of the AI algorithms.