With the development of power electronics technology, the AC/DC hybrid distribution network based on power electronic transformer (PET) can better accommodate distributed power sources and DC loads, but it also presents obvious zoning characteristics and stronger operational uncertainty in the distribution network. Firstly, this article constructs a three port steady-state model of PET based on the demand for flexible power regulation in AC/DC hybrid distribution networks. Secondly, a method combining cloud modeling, non parametric kernel density estimation (KDE), and confidence intervals was proposed to model the uncertainty of wind and photovoltaic output power prediction errors. Finally, a day-ahead optimal scheduling model of AC/DC hybrid distribution network containing PET was constructed and solved, and the validity of the proposed model and method was verified by numerical examples.