Landslides, which are aggravated by climate change, greatly threaten mountainous regions like northern Pakistan. However, existing research lacks a complete, region-specific analysis of the climatic and environmental factors driving landslides across various climatic zones, specifically in vulnerable areas such as northern Pakistan. This study explores the N-15 Highway in northern Pakistan. This region is frequently impacted by landslides induced by extreme climatic events, including heavy rainfall and flooding, which usually lead to blockages along the route. We collected a complete landslide inventory using 455 satellite images from 1990 to 2023 and ground surveys. We also analysed the relationship between landslides and climate change over the period of 1990-2023, encompassing soil moisture, vegetation, precipitation, temperature and snow cover. Using meteorological data, we found that the frequency of landslides rose exponentially from 1990 to 2023 due to the impacts of climate change. Especially after 2005, substantial increases in precipitation, temperature and snowmelt led to a more significant rise in landslide occurrences (p < 0.05). In the warm season (April-October), 84.1% of the landslides occurred, which were mainly due to precipitation and snowmelt. Balakot, Babusar-Naran and Chilas were the primary areas along the highway, each with distinct landslide mechanisms. In the Balakot region, which is characterised by sub-tropical conditions, high precipitation played the leading role in landslide occurrences. Landslides at Babusar-Naran, which is known for Alpine conditions, were mostly driven by precipitation, soil moisture fluctuations and snowmelt dynamics. Geological reasons and high temperatures influenced the Chilas region, which is characterised by semi-arid conditions. The EC-Earth3 model from CMIP6 predicts a 1.6-6.5 degrees C warming and a 35% rise in precipitation by 2100, with more extreme variations under SSP3-7.0 and SSP5-8.5 scenarios. These changes are likely to result in arise in the frequency of landslides. We suggest improving ground observation networks and utilising multiple datasets to better understand the relationship between landslides and climatic variables, which enables highly accurate risk assessment and management in high-mountain areas under the warming climate.