Promoting the cultivation of computational thinking (CT) skills in programming activities has become a key issue in the K-12 curriculum, however, there is no unified conclusion on how to design programming activities to promote the acquisition of CT skills more effectively. The purpose of this study is to determine the effectiveness of using programming to enhance K-12 students' CT skills and explore the influence of various programming instructional design factors on the acquisition of CT skills. This study presents a quantitative meta-analysis. A systematic search of randomized controlled studies on the influence of programming on CT skills. A total of 86 empirical studies with 114 effect sizes met the study selection criterion. The results showed that programming improved K-12 students' CT skills in general (Hedges' "g" = 0.601, 95% CI [0.505, 0.697], p < 0.001). No publication bias was detected. Besides, we found that the interdisciplinary integration of programming, the duration of programming intervention within 1 week to 1 month, the class size of less than 50 students, and a reasonable selection of programming instruments and CT assessment types may be more conducive to promoting students' CT skills. Based on these findings, we suggest that programming instructional design should be planned reasonably to make the best use of technology towards developing students' CT skills. This will be of great significance to programming teaching and CT education in K-12.