Intelligent detection of polluted gas leakage from pipelines in chemical parks can replace manual labor. However, the coordinated use of multiple data in sensor-based methods remains a difficult problem. The proposed method is based on the distributed fiber optic sensing data fusion algorithm for detecting the leakage of polluted gas from pipelines in chemical parks. To identify detection indices for characterizing polluted gas leakage in pipelines within chemical parks, we selected appropriate fiber optic temperature sensors, fiber optic pulse sensors, and fiber optic strain sensors to build a distributed fiber optic sensor network. We collect the monitoring indexes at the leakage detection points of polluted gases from pipelines in chemical parks. The pulse data are processed by using the flow balancing algorithm, and the temperature and strain data are processed by using the scalar Kalman filtering algorithm. The weighted fusion algorithm is used to find the corresponding weights adaptively according to the measured values of each fiber optic sensor. Based on these weights, the results of the above-mentioned data processing are fused. The fusion results are sorted according to the time series to determine whether the polluted gas leakage from pipelines in the chemical park is indicated by a sudden change in waveforms. The results show that the method can accurately collect the relevant information on the detection indexes and realize the accurate detection of pipeline gas leakage accordingly, effectively reducing the environmental pollution and safety hazards caused by leakage.