Research has found that levels of enjoyment and anxiety fluctuate on a moment-to-moment timescale during second language (L2) spoken task performances as learners attempt to cope with various communication challenges. For L2 speakers, surges in these emotions can limit or expand cognitive resources, affecting speech processing capability. However, at an intra-individual level, there is very little empirical evidence on how emotions and fluency are related during L2 spoken task performances. The present study uses the idiodynamic approach to examine the relationship between enjoyment, anxiety, and breakdown fluency (i.e., average length of pauses) during monolog tasks performed by university students who use English as an L2. After watching a video recording of their task performances, participants rated their anxiety and emotion levels on a per-second basis. Immediately after, they were interviewed about their attributions for fluctuations in their ratings. After segmenting task performances into 26 7-s segments of speech, per-person correlations revealed that (1) the (negative) relationship between anxiety and enjoyment varied from strong to very weak, and (2) the (positive) relationship between anxiety and breakdown fluency was much stronger than the (negative) relationship between enjoyment and breakdown fluency. Triangulation of anxiety and enjoyment ratings, stimulated recall interviews, and performance data led to the identification of four categories of factors that influenced the emotion-fluency relationship: task design factors (task structure, task topic), task implementation factors (pre-task planning, task time limits), cognitive-linguistic factors (momentary breakdowns in conceptualization and formulation speech processes), and achievement outcome factors (self-evaluations of appropriate and inappropriate language used).