The method to produce accurate and precise age-at-death estimates from skeletal remains is crucial for both forensic and archaeological analyses. Despite decades of efforts, all the currently available aging methods are experience-based and unable to generate satisfactory results. The Transition Analysis (TA) developed by Boldsen and his colleagues is an advanced method of estimating age from skeletons based on Bayesian statistics. It can combine multiple skeletal indicators of age, avoid the effect of age mimicry, and allows the ages of elderly people to be estimated more precisely. Although TA has been well received by foreign researchers, we still knew little about this method and seldom adopted it in China. The Yuejiazhuang cemetery is located in Luochuan, Shaanxi Province, and dated to from the mid-Warring States Period to the unification of the Qin Dynasty. It’s the northernmost large public cemetery of Qin people ever found in the Shaanxi region. Altogether 304 skeletons were recovered during the archaeological excavation in 2020. We used both the TA and the traditional methods to estimate the ages of skeletons excavated from the Yuejiazhuang cemetery and made a comparison of the results. The data showed that both age-at-death distributions and survival processes are remarkably different. The maximum lifespan and the average age of death obtained by the TA method are much higher, thus indicating that more adult individuals could survive into their middle and old age. These two methods generate completely different age structures of death, which will definitely affect our understanding of the past, such as people’s age perception, intergeneration relationships, family size, and social structure. Besides, the target sample’s preservation and age structure also have a great impact on the age differences estimated by the two methods. When the skeletons are poorly preserved or the individuals are relatively young, the age differences between the two methods become smaller. In conclusion, the TA method represents the most recent and significant development in the skeletal age estimation approaches, which standardizes the observation process and quantifies errors. However, TA is not yet perfect and still needs to be modified. All kinds of TA software so far provide only one algorithm and users can’t choose mathematical models according to their requirements. Wide age intervals and systematic age-estimation bias still exist, particularly for individuals in the youngest and oldest portions of adulthood. No human bones from China are included in the reference sample of TA. Therefore, more validation studies using known-age skeletal collections are needed in the future to evaluate TA’s performance on Chinese samples.