Micro-grids provide opportunities to enhance reliability and flexibility in power systems. In recent years, resiliency of power grids has attracted many research interests. Moreover, the optimal energy scheduling of micro-grids can satisfy resiliency improvement of the system. In this paper, a novel two-stage risk constrained stochastic framework is proposed, which can optimally schedule a dependent micro-grid in both normal and emergency situations. Dependent micro-grid is introduced as a new group of micro grids with additional interconnection point. It can use energy storage systems, DGs and demand response resources to guarantee resilient and economic operation. In the proposed model, linearized AC grid constraints are added to the mathematical formulation to improve computational efficiency. Predominant uncertainties such as loads, prices, wind, and unplanned islanding events are managed based on risk constraints. The optimal energy management strategy is formulated as an MILP and implemented in the IBM CPLEX (R) Software. Several case studies are carried out to demonstrate the application of the proposed framework. The sensitivity of energy storage system scheduling and the operator's participation in the electricity market are examined by adjusting risk parameters. Furthermore, it is shown that energy procurement by an extra interconnection point would mitigate the economic losses during emergency condition. (C) 2019 Elsevier Ltd. All rights reserved.