Climate variability and change pose serious challenges to smallholder farmers and agro-pastoralists. Nonetheless, their trends and impacts on rain-fed farming system (RFFS) in semi-arid areas of Tanzania have not sufficiently been explored. This study was conducted in Iramba and Meatu districts to contribute to this knowledge gap. The study specifically (i) assessed meteorological data trends of rainfall and temperature between 1994 and 2011; (ii) determined farmers’ perception of climate variability and change in relation to meteorological data trends; (iii) determined changes in RFFS in response to climate variability and change; and (iv) examined changes in gender relations in response to climate variability and change. Although available meteorological data were less than 30 years suggesting climate variability, farmers’ perceptions covered up to 30 years and so addressed the question of climate change. A qualitative phase informed a household survey that covered a random sample of 388 households’ respondents (39% women). Qualitative data were transcribed into text and analyzed based on content and meaning of the text. Statistical Package for Social Sciences (SPSS) was used to analyze quantitative data. The results showed that there was no significant increase (P > 0.05) in inter-annual rainfall variability. However, seasonal rainfall variability showed a clear decreasing trend in April and December, in Iramba; and in January and April, in Meatu. Decreasing rainfall trend in April occurred simultaneously with increasing temperature trend. In addition, farmers’ perception and meteorological data trends compared well on change and on increased rainfall unpredictability as well as on increased warming and dry years. Nonetheless, due to missing data in some periods, meteorological data trends did not show increased frequency of drought since the 2000s as opposed to farmers’ perception. As hypothesized (P > 0.05), men and women’s perceptions were almost the same. Similarly, perceptions of the poor, not so poor and the rich were almost the same (P > 0.05). Unlike the hypothesis, the binary logistic regression model showed that climate variability and change had significant impact on changing crop varieties and livestock grazing places relative to non-climatic factors. Warming (ß = -10.61, Wald = 36.26, P ≤ 0.001) showed the highest impact on changing crop varieties. In addition, drought (ß = 2.16, Wald = 6.82, P ≤ 0.009) showed the highest impact on changing livestock grazing places. Based on division of labour, control over resources and biased norms, climate variability and change increased and perpetuated existing asymmetrical gender relations. The study concludes that although inter-annual rainfall had not changed significantly, temperature, drought and seasonal rainfall variability had intensified relative to the situation in the 1970s. This had substantial impacts on cropping and livestock systems and on gender relations. Therefore, strategies used by the farmers and interventions promoted by the government and Non-governmental Organizations (NGOs) to address the impacts should comprehensively consider seasonal variability with gender perspectives.Deutscher Akademischer Austausch Dienst (DAAD,(EPINAV)and Higher Education Institutions Institutional Cooperation Instrument (HEI-ICI.