There is an increasing demand for reliable methods to measure soil loss in order to get a better understanding of sediment movement and to develop process‐based sediment transport models. This study described the design, calibration, and practical application of an instrumented technique that automatically characterized surface flow in situ fields. The instrument consisted of a sampling assembly, a measuring assembly, and a management platform. The special design combined with a volume‐mass conversion relationship was used to determine the exact volume and mass of the sediment‐laden water, thereby allowing the implementation of real‐time measurement and synchronous calculation of runoff rate and sediment concentration. Calibration results showed that normal distributions and an average error of −3.67% were obtained for the target sediment concentrations of 2–1,000 g/L. Compared with the gravimetric method, the average error of the new technique for sediment concentration was 4.82%, and the highest precision was 98.50%. The high repeatability and accuracy of the technique suggest that it can be used with confidence to give satisfactory results in fields. A field experiment under natural rainfall showed that a grass‐covered plot had 18.94% less runoff rate and 42.92% lower sediment concentration than a bare soil plot, illustrating the impact of vegetation cover on reducing soil loss. This uniquely designed instrument provides a new approach to real‐time sediment detection in surface flow. It will be beneficial for rigorously evaluating runoff and sedimentation and for in‐depth research regarding water and soil loss processes.