QTL mapping using F1 population is an important strategy to study the quantitative traits of the tea plant. However the results are largely influenced by the size of population. In this study, sixty F1 populations with different sizes (n = 50, 100, 150, 200, 250, 300) were generated by randomly sampling from an original population (N = 327), which was derived from a cross of ‘Longjing43’ × ‘Baihaozao’. By integrating a SSR-based genetic map and phenotype data of three traits (timing of spring bud flush, young shoot color and leaf shape index), QTL mapping was performed in all populations. The results showed that the size of population results in a significant influence on estimation of phenotypic variation explained (PVE). When 150 F1 individuals were applied, QTLs with an effect of >10% were found. Notably, the estimation of QTL effect value remained un-precise, since 2.8-fold difference on estimated effect was observed between highest and lowest values. Therefore, a larger size of population is suggested for precisely effect estimation, and QTLs identified from a small F1 population can be employed with caution in the marker-assisted breeding.