Small leaks, many of which often go undetected using conventional gauges, remain an urgent problem for an aging pipeline infrastructure. This article proposes a method that allows for an automated and robust leak detection and localization system by fusing distributed acoustic sensor (DAS) and distributed temperature sensor (DTS) data for machine learning. Distributed fiber-optic sensing creates an advantage over conventional gauges by providing real-time continuous measurements along the entire length of the fiber-instrumented pipeline at a high spatiotemporal resolution and sensitivity that can detect small leaks, as well as the leak location. High sensitivity, however, can create noisy data. Thus, machine learning is applied for a robust method of distinguishing nonleak data from leak signatures for accurate leak detection and localization. The workflow is demonstrated on an experimental pipeline setup exposed to environmental noise. The results illustrate reliable detection of small leaks between 0.04 and 0.30 L/s with F1 scores over 0.9, on a range of DAS frequency bands combined with DTS data. The fusion of two different types of distributed fiber-optic sensors not only increases the likelihood of small leaks being detected, but also decreases the number of false alarms by filtering noise that only appears in one sensor domain. Furthermore, the machine-learning approach utilizes segmentation, allowing for precise localization and quick investigation, as well as AI explainability of the sensor data deemed a leak signature.