After the smart home system is impacted by malicious attacks or other factors, the system reliability drops drastically, so the user's living environment is greatly disturbed. To determine the emergency maintenance strategy, firstly, we study the failure mechanism of the smart home system by the fault tree theory, and distinguish the components into two types: critical components and non-critical components. Secondly, a recovery prioritization is proposed to measure the extent to which changes in the reliability of each component affect the reliability of the system. Then, three scenarios of smart home system failure are proposed. Under the constraints of time and cost, the maintenance strategies for the three scenarios are determined to achieve emergency recovery. Finally, through the multi-objective particle swarm optimization algorithm, the system reliability is maximized, the maintenance cost is minimized, and the optimal reliability recovery level of each component is obtained, so that the reliability of the smart home system is further improved. The effectiveness and practicality of the emergency maintenance strategies proposed in this paper are demonstrated by a smart home system example.