Particle swarm optimizer (PSO) is a population-based optimization technique applied to a wide range of problems. In the literature, many PSO variants have been proposed to deal with noise-free or noisy environments, respectively. While in real-life applications, noise emerges irregularly and unpredictably. As a result, PSO for a noise-free environment loses its accuracy when noise exists, while PSO for a noisy environment wastes its resampling resource when noise does not exist. To handle such scenario, a PSO variant that can work well in both noise-free and noisy environments is required, which does, to the authors' best knowledge, not exist yet. To fill such gap, this work proposes a novel PSO variant named as dual-environmental PSO (DEPSO). It uses a weighted search center based on top- k elite particles to guide the swarm. It averages their positions rather than resampling fitness values of particles to achieve noise reduction, which challenges the indispensable role of the resampling method in a noisy environment and adapts to a noise-free environment as well. Two theoretical analyses are presented for noise reduction and finer local optimization capabilities. Experimental results performed on CEC2013 benchmark functions indicate that DEPSO outperforms state-of-the-art PSO variants in both noise-free and noisy environments.