The detection of small side branches within underground water pipe networks can be challenging since they are usually subject to many unknown uncertainties. A transient-based method was proposed to detect multi-branching pipes in pipeline systems having uncertainties associated with unknown abnormalities. To isolate the uncertainty signal from the original pressure signal, a polynomial transient model was developed that can address nonlinear transient introduction. Through the superimposition of different pressure signals, the uncertainty signal was delineated to obtain the processed signal for multi-branch prediction. Pipeline systems with two branching pipes were tested under various uncertainties to detect multi-branching pipes. In this study, two transient generation conditions were used: an instant valve closure and a pressure injection. Under several uncertainty conditions, hydraulic pressure damping can be obtained for the detection of two branching pipes under both transient generation conditions. It is possible to predict the locations of multi-branching pipes with a prediction error of less than 10% by using proper signal processing and manipulation for two distinct transient generation methods in a pilot-scale pipeline system with several uncertainties. This study demonstrates the potential for detecting multi-branching pipes in a pipeline system, even with unknown uncertainties for a pipeline system.