Cognitive style represents individuals’ consistence in information acquisition and information processing. It affects individual’s cognitive functioning, such as memory, attention, and problem-solving. To describe the developmental trend of children’s cognitive style and to investigate how cognitive style affects children’s problem-solving behaviors, we measured 98 4- to 8-year-old children’s wholistic-analytic dimension of cognitive style and the strategies they used in problem-solving. Participants were categorized into three age groups: preschoolers of the middle class and the senior class, as well as first-grade primary students. A computer program-based Cognitive Style Analysis (CSA) test was used to measure participants’ cognitive style. The stimuli were presented in a ThinkPad Yoga laptop with a resolution of 1920×1080. Participants were required to judge whether two figures were identical in the wholistic subtask and whether the simple figure was embedded in the complicated figure in the analytic subtask. Task order was counterbalanced among participants. Cube-based Puzzle Game was used to measure the participants’ problem-solving strategies. In the task, participants were required to build the same pattern as in the task card. First, regarding the cognitive style analysis test, we found that primary school children had higher accuracy and shorter reaction time in both the wholistic and the analytic subtasks than preschoolers, which indicates that cognitive ability grows with age rapidly. However, there was no significant difference between preschoolers of the middle class and the senior class. Second, we compared the cognitive style (wholistic-analytic ratio, calculated as the performance on the wholistic task divided by the performance on the analytic task) among children of the three age groups. One-way ANOVA showed that there was no significant age differences, and the average wholistic-analytic ratio was 1.84, suggesting that children aged 4 to 8 overall tended to be wholistic. Meanwhile, the analytic style gradually develops with age, as Pearson correlation results showed that age was significantly negatively correlated with the wholistic-analytic ratio. Finally, we divided participants into the high and low ratio group according to the median of the wholistic-analytic ratio. Mann-Whitney U Test results showed that the high ratio group (relatively wholistic style) used Subject Reference strategy (first complete the main part and then the background) more frequently than low ratio group (relatively analytic style), while low ratio group preferred Clue Inference strategy (complete the puzzle in the order of rows or columns strictly) and Local Positioning strategy( place the cube where it should be without obvious sequence). This study is one of the first research to explore how cognitive style develops in young children and how it influences the problem-solving strategy in early childhood. We found that 4- to 8-year-old children tended to be wholistic cognitive style; meanwhile, the analytic cognitive style increased with age. Importantly, cognitive style influenced children’s problem-solving strategies: children with relatively wholistic cognitive style tended to use Subject Reference strategy in problem-solving, while children with relatively analytic cognitive style tended to use Clue Inference strategy and Local Positioning strategy. The research facilitates our understanding of children's development of cognitive style and its effect on problem-solving behaviors. It may also provide theoretical evidence for early childhood parenting practices and education for children’s problem-solving.