Energy transactions among multiple microgrids (MMGs) can improve economy and reliability of electricity markets. However, power flow constraints (e.g., voltage tolerance and node balance constraints) with respect to distribution networks (DNs) are usually overlooked for reducing the complexity of energy transaction problems in existing studies. This paper models a bi-level energy trading framework between MMGs and DNs. Firstly, in order to guarantee the safety operation while minimizing the total cost of DNs, full alternative current power flow constraints are considered at the upper level. Then a comprehensive energy trading model for MMGs is established at the lower level. Furthermore, analytical target cascading (ATC) algorithm is applied to solve energy trading problems between MMGs and DNs in a distributed manner. Finally, numerical tests on an IEEE 33-bus distribution system are implemented to demonstrate the performance of our distributed energy transactions method.