Abstract: It is difficult to accurately obtain the renewable energy output and load based on the prediction technology. The randomness has a profound impact on the control strategy of micro-grid and the economy of users. For the grid connected micro-grid with photovoltaic system, energy storage system and load, a micro-grid day-ahead optimal scheduling model with demand management and bundled with peak-valley arbitrage is established, the objective function of which is the lowest daily electricity payment. Based on this model, taking into account the fluctuation of photovoltaic output and load, a multi-objective robust optimal scheduling model of micro-grid based on information gap decision theory (IGDT) is proposed to provide robust strategies, and study the quantitative relationship between preset target cost and photovoltaic output and load fluctuation range. The ε-constraint method is proposed to depict Pareto efficient frontier of multi-objective problem. Besides, the fuzzy satisfaction theory is used to determine the compromise solution of Pareto solution set, which provides a reasonable decision-making basis for operators. Compared with Monte Carlo method, simulation results on the micro-grid in an industrial park verify that the proposed model is feasible and effective.