This paper combines panel data from nationally representative household-level surveys in Nigeria with long-term satellite-based spatial data on temperature and precipitation using geo-referenced information related to households. It aims to quantify the impacts of climate change on agricultural productivity, income shares, crop mix, and input use decisions. We measure climate change in harmful degree days, growing degree days, and changes in precipitation using long-term (30 year) changes in temperature and precipitation anomalies during the crop calendars. We find that, controlling for other factors, a 15 percent (one standard deviation) increase in change in harmful degree days leads to a decrease in agricultural productivity of 5.22 percent on average. Similarly, precipitation change has resulted in a significant and negative impact on agricultural productivity. Our results further show that the change in harmful degree days decreases the income share from crops and nonfarm self-employment, while it increases the income share from livestock and wage employment. Examining possible transmission channels for this effect, we find that farmers change their crop mix and input use to respond to climate changes, for instance reducing fertilizer use and seed purchases as a response to increases in extreme heat. Based on our findings, we suggest policy interventions that incentivize adoption of climate-resilient agriculture, such as small-scale irrigation and livelihood diversification. We also propose targeted pro-poor interventions, such as low-cost financing options for improving smallholders’ access to climate-proof agricultural inputs and technologies, and policy measures to reduce the inequality of access to livelihood capital such as land and other productive assets.