The invention relates to a perception layer data fidelity method for a local area internet of things. The perception layer data fidelity method comprises the steps that a trusted execution environmentprocessor is arranged between a sensing node and a gateway in the local area Internet of Things, and the trusted execution environment processor can be used for intercepting a risk sensing node witha suspected identity before sensing data collected by the sensing node passes through the gateway, wherein the trusted execution environment processor is configured to extract an identity attribute set of the sensing node before the sensing node accesses the gateway, and the identity attribute set is used for identifying the identity of the sensing node. The identity authentication of the sensingnode is carried out in a trusted execution environment, and the trusted execution environment is a completely trusted part in the network and can be responsible for sensitive operations such as identity authentication, so that the identity authentication of the sensing node is completely trusted in the trusted execution environment.