<p>Vegetation dynamics in northern Nigeria are strongly influenced by climate variability; however, long-term interactions remain underexplored. In this study, NDVI data from GIMMS AVHRR (1981–2015) and MODIS Terra (2000–2021) and daily precipitation and temperature records from ten synoptic stations (1980–2021) across major ecological zones were analysed. The NDVI datasets were harmonized and georeferenced, and climate extreme indices were computed using RClimDex following the ETCCDI guidelines. Vegetation trends and variability were assessed using the Mann–Kendall test and coefficient of variation. Long short-term memory (LSTM) models with multiple lag periods were employed to simulate the responses of the NDVI to climate extremes, and model performance was evaluated using R² and RMSE. The results indicate that the vegetation in the Sudan Savanna Basin exhibited a modest positive NDVI trend with high interannual variability, whereas that in the Guinea Savanna Basin experienced a long-term decline from 0.26 in 1980 to 0.16 in 2021, despite intermittent recoveries. The Sahel Savanna displayed strong interannual fluctuations but an overall neutral trend, and the tropical rainforest maintained relatively stable cyclical NDVI variations. Among the LSTM models, the 12-month lag model performed best (training R² = 0.846; testing R² = 0.837; RMSE = 0.26), suggesting a delayed vegetation response of approximately one year to climate extremes. SHAP analysis revealed total precipitation and maximum 5-day rainfall as the primary drivers of NDVI variability, with temperature extremes exerting a weaker influence. The results of this study reveal that precipitation plays a critical role in regulating vegetation health. To sustain ecosystem productivity, targeted climate adaptation and land management strategies are recommended for northern Nigeria.</p>