The observer-based stabilization of network-based interval type-2 (IT2) semi-Markov jump models is investigated, where a new control strategy based on redundant channels is proposed to overcome the adverse effects of packet loss. The IT2 fuzzy method is adopted to describe the nonlinear objects, which overcomes the uncertainty problem of the traditional T-S fuzzy model. An observer-based control strategy is adopted due to the difficulty of obtaining complete state information under a complicated network environment. The main novelty is that redundant channels are adopted to construct a suitable observer-based feedback control scheme, and the auxiliary variable is introduced to solve the matrix dimension problem to achieve better dynamic performance of IT2 fuzzy networked semi-Markov jump models, improve the stability and anti-interference ability of the system, and overcome the difficulty caused by potential packet loss. Based on the semi-Markov kernel, IT2 fuzzy, and the Lyapunov functions dependent on the elapsed time, sufficient criteria are constructed for the considered system to ensure the $\sigma$ -error mean-square stability. Furthermore, the expected solvability is constructed for an observer-based feedback controller under a linear matrix inequality framework. Ultimately, the tunnel circuit model verifies the proposed control method.