Planning is an important developmental outcomes in middle childhood, and also is an important correlate of academic achievement during school years. Planning is defined as the ability to formulate a strategy prior to action, execute the strategy, and verify whether the strategy is effective for solving a problem. It was argued that the planning process involved different attention skills, such as sustained attention to focus on the task, selective attention on the target characteristics, and attentional control to inhibit the inappropriate but preponderant responses. The previous studies showed attentional control and sustained attention significantly correlated with the development of planning, but it is unknown whether selective attention uniquely predicts planning. In addition, most of these studies are not longitudinal, and thus little is unknown about the predictive power of earlier attention skills to later planning and also the direction of the relationships between planning and attentions skills. Therefore, this study aims to examine the dynamic developmental relationships between planning and different components of attention skills (sustained attention, attention control, and selective attention) in middle childhood. One hundred thirty-nine 7-year-old children from a primary school in Shanghai city (67 boys and 72 girls; mean age = 89.81 months, SD = 3.22 months) were followed until 9 years old, and were annually assessed on planning, sustained attention, attentional control, and selective attention. Planning tasks were from the Planning subtest of Das-Naglieri Cognitive Assessment and attention were assessed by tasks from Test of Everyday Attention for Children. Pearson Correlations among planning and three components of attention (composite scores) indicated a significant stability for these four skills during the period from age 7 to age 9, and significant correlations among these variables at three points. The cross-lagged model was used to examined the developmental relationships between planning and three attention skills, and fitted well to the data (χ2/df = 2.30, GFI = .96, CFI = .97, RMSEA = .09). The results of crossed-lagged analyses showed the effects between planning and attentional and these between planning and selective attention were bidirectional during the period from age 8 and age 9, but sustained attention and attentional control at age 7 predicted planning at age 8, and planning at age 7 predicted selective attention. The results indicated a mutual relationship between planning and different attention skills.