In various disasters, fire is the most major disasters, threats to public safety and social development. The highest frequency of occurrence of the fire in the room, for fire monitoring focuses on environmental information (temperature, fire, smoke, etc.) monitoring, so these three parameters can monitor and alarm fire accurately. However, in the conventional system does not consider whether the sensor data obtained credible, this paper plausibility calculation for the obtain data, which excluded outliers, and credible data weighting, finally use the merge weighting parameters get the result.