<p>Today, many real-world optimization problems are characterized by multiple objectives, constraints, and parameters that change over time, and these are often called dynamic multiobjective optimization problems (DMOPs). A significant challenge faced by traditional Evolutionary Algorithms (EAs) in addressing DMOPs lies in efficiently and accurately tracking the shifting Pareto-optimal front and maintaining diverse, high-quality solutions throughout the optimization process. Transfer learning (TL)-based methods have the ability to reuse the knowledge gained from previous computed solutions to enhance the quality of current solutions. However, existing transfer learning (TL)-based approaches require a lot of computing resources or cannot efficiently track the true Pareto front in fast-changing environments. This paper proposes a method to solve DMOPs using an archive-based Transfer Component Analysis (TCA) framework (ATCA-DMOEA). TCA minimizes the distance between the previously mapped optima and the current problem domain, integrating it with a multiobjective EA (MOEA), called Nondominated Sorting Genetic Algorithm-II (NSGA-II). The method stores the best individuals from past computations in an archive with the transfer learning (TL) feature to predict the optimal individuals at the new instance during the evolution. Using this strategy, we do not need to build or train a model from scratch, which reduces the computational cost required by existing methods. Different benchmark problems are used to validate the proposed algorithm, and the simulation results are compared with various state-of-the-art dynamic multiobjective optimization algorithms (DMOAs). The results show that our approach can improve computational speed while achieving better quality solutions than existing methods.</p>