With the increase in customer customization demands and the focus on delivery times, just-in-time production has become a key factor in enhancing enterprise competitiveness, making job shop scheduling for just-in-time production worthy of in-depth study. To address the issues of severe delays and low just-in-time performance in job shop scheduling, a job shop scheduling model is proposed with the objectives of minimizing delays, minimizing lead times, and minimizing the maximum completion time. To solve this model, a self-learning hybrid CHIO algorithm (HCHIO) based on the coronavirus herd immunity optimizer (CHIO) is proposed. Firstly, a self-learning operator library with a scoring evaluation mechanism is designed, enabling the algorithm to self-learn and select the optimal operator for different problems to enhance the global optimization performance of the algorithm. Secondly, by conducting neighborhood search on the optimal solution, the local search ability of the algorithm is strengthened. Experiments on benchmark tests and real cases were conducted on HCHIO, verifying the excellent optimization ability of the proposed algorithm in solving job shop scheduling problems. The experimental results demonstrate the effectiveness of the self-learning hybrid CHIO algorithm in solving just-in-time job shop scheduling problems.