Abstract: Beijing, a prominent megacity in China, experiences notable interannual variation in its seasonal precipitation cycle, which exerts substantial influences on both economic activities and societal well-being. Based on daily precipitation observations collected at 20 meteorological stations in Beijing over the period from 1981 to 2022, this study reveals spatiotemporal variability characteristics of local precipitation seasonal cycle anomalies and elucidates their underlying causes using self-organizing mapping neural networks and statistical diagnostic methods. The seasonal cycle of precipitation anomalies during the flood season in Beijing manifests four distinct types. Notably, two archetypal features, namely "persistent drought or flood" and "rapid change of drought and flood", are directly linked to circulation anomalies with the barotropic structure over the Sea of Japan. These two types account for 45.2% and 54.8% of the total samples, respectively. In years characterized by persistent droughts or floods, the occurrence of a high-pressure (low-pressure) anomaly over the Sea of Japan triggers a coupling of mid-latitude Mongolian cyclones and the South Asian high. Consequently, a low-level northerly (southerly) anomaly prevails in the local area, inhibiting (facilitating) moisture transport from the Northwest Pacific and resulting in sustained deficient (sufficient) precipitation. In the years featuring rapid changes in drought and flood, when an anomalous low (high) pressure system prevails over the Sea of Japan in July and the Western Pacific Subtropical High anomalously shifts southward (northward), precipitation in Beijing region significantly reduces (increases). In August, with the abnormal strengthening (weakening) of the Mongolian cyclone and the transition to an anomalous high (low) pressure system over the Sea of Japan, precipitation in Beijing region increases (decreases) subsequently. Furthermore, persistent drought- or flood-type anomalies exhibit a significant correlation with extratropical sea surface temperature. In contrast, drought-flood rapid transition anomalies are strongly associated not only with tropical El Niño-Southern Oscillation (ENSO) events but also with 10—30-day intraseasonal oscillation of local precipitation.