Partial blockages in long-distance natural gas transmission pipelines lead to a poor operating efficiency for the entire system and increase the risk of a potential accident. Therefore, effective detection and real-time monitoring of partial blockages are important to ensure safe operation of a natural gas pipeline. Previous studies have primarily examined partial or complete blockage detection for natural gas pipelines under shutdown conditions or techniques that require a reduction in the pipeline pressure or flow before engaging in related detection operations. However, as natural gas pipelines are a closed-pressure conveying system, reducing the throughput of a pipeline or shutting down a pipeline directly affects the upstream and downstream users, which results in economic loss to both the pipeline operators and consumers. Based on an analytical solution method previously proposed for gas transient flows in a pipeline containing a partial blockage, this paper presents a theoretical partial blockage detection method for natural gas pipelines under normal operation conditions. First, we used a partial blockage model and its analytical solution as the direct partial blockage parameter inversion problem and calculated the mass flow rates plus random errors to be used as the observed values. Afterward, we used the Tikhonov regularization method to establish the optimization objective function, which contained the partial blockage parameters and random measurement errors. A genetic algorithm is adopted to solve the inverse problem of parameter identification. Multiple groups of numerical experiments indicate that the proposed theoretical method can be used for natural gas pipeline extended partial blockage parameter identification for a pipeline under normal operational conditions and that the relevant observed parameters contain random errors.