Lake Merzbacher, located in the central Tian Shan Mountains, experiences frequent outbursts. The resulting glacial lake outburst floods (GLOFs) pose a severe threat to the water resource security of the downstream basin. To address the gaps in existing research regarding the response mechanisms and early warning indicators of GLOFs, this study integrated multi-source remote sensing imagery, meteorological reanalysis data, and runoff observation data from hydrological stations spanning 1990 to 2022. Utilizing a combination of the NDWI and visual interpretation, we constructed a high-precision, high-temporal-resolution dataset of glacial lake boundaries over the past three decades and extracted GLOF events to analyze the evolutionary characteristics of the glacial lake and its outburst floods. The results reveal that the evolution of the glacial lake area can be divided into three stages: a rapid expansion phase (1990–1996), a relatively stable phase (1997–2010), and a structural transformation phase (2011–2022). GLOF events predominantly occurred in July and August (accounting for 96% of occurrences) and exhibited an advancing trend in timing (7 days/decade). The peak discharge demonstrated an increasing trend (approximately 66 m3/s per decade) with heightened abruptness (flood duration shortened by 1.1 days/decade), and the overall flood magnitude has intensified since 2010. Furthermore, we quantitatively analyzed the relationships among the maximum pre-outburst area, lake drainage volume, net peak discharge, and total flood volume. By incorporating the pre-outburst glacial lake area, floating ice coverage, and changes in the altitude of the 0°C level height, we proposed a set of multi-dimensional early warning indicators capable of evaluating outburst timing and flood magnitude. These findings provide a scientific basis and reference for enhancing forecasting and early warning capabilities, as well as disaster risk assessment for GLOFs.