The paper introduced the concept of a multilevel microgrid (MG), which starts at the building level and creates an MG at this level. The development process of MG in China was described in detail for the first time, including the key issues and the research direction. The prediction error and uncertainty analysis of wind power, photovoltaic power generation, load, and controllable generator components are given, respectively. The uncertainty distribution model is helpful to construct the objective function of minimizing the operating cost of MG. First, an hypothetical apartment building's suitability for installation of solar panels, wind turbines, and solar thermal devices in Shanghai was analyzed, including the principles and characteristics of their electrical generation and energy collection. Compared to the analysis of a single building, the results are more comprehensive for simulating the loading of different buildings in a variety of situations. The mathematical model of peak load shaving in an MG was constructed, and the interior-point method was used to solve it. By simulation analysis, the variation of load curve and related parameters before and after peak load shaving was obtained. First, the average loading of an office building with and without the MG was compared, and the reduction rate in peak loading of a building was used to analyze whether there was additional storage and found the change rule. A neighborhood with the additional MG technologies and additional storage units is better than other existing literature in showing the effectiveness of peak- load reduction.