Earthquakes in the Himalayan region primarily result from the ongoing convergence of the Eurasian and Indian tectonic plates, with the Indian plate being underthrust beneath the Eurasian plate, leading to complex seismic activity. The Nepal region, located within a seismic gap, is especially significant due to its potential for large earthquakes. In this study, we investigate the temporal evolution of seismicity in the Himalayan region to identify precursor signals preceding the 2015 Mw 7.8 earthquake in Nepal Himalaya, with a focus on the fractal characteristics of seismic events. The magnitude of completeness (Mc) for the dataset was found to be 3.9 and above. The sliding window analysis using 100-event windows showed variation in fractal correlation dimension (Dc) values over time. During the first nine years, the trend increases; later, it significantly decreases. Increasing and decreasing trends demarcate the two time periods during which the fractal behaviour of earthquake occurrences differs. We detected unusual patterns in the fractal correlation dimension (Dc) of seismic events, with values ranging from 0.94 to 1.66. A sinusoidal-like pattern emerged in the initial years of the dataset (within the first 9 years), followed by a notable change in the fractal dimension (after 9 years), just prior to the main earthquake. This indicates a shift towards stronger clustering of earthquakes as stress built up in the region. The patterns in the fractal dimension suggest significant changes in seismic behaviour before 11 years leading up to the 2015 earthquake, indicating stress accumulation and the likelihood of a large seismic event. Notably, the time period of the decreasing trend exhibits a comparatively higher crustal stress than the period of the increasing trend. This study provides insights into long-term stress accumulation patterns and their application in seismic hazard assessments.