Increasing evidence in humans and other animals suggests that testosterone (T) plays an important role in modulating emotion. We previously reported that T treatment in rhesus monkeys undergoing chemically induced hypogonadism results in increased watching time of videos depicting fights between unfamiliar conspecifics (Lacreuse et al., 2010). In the current study, we aimed to further investigate the effect of T manipulations on attention and memory for emotional stimuli in male rhesus monkeys. Six males (7 years old) were administered Depot Lupron to suppress endogenous T levels and treated with either testosterone enanthate (TE, 5 mg/kg) or oil, before crossing over to the alternate treatment. Animals were tested for 16 weeks on two computerized touchscreen tasks with both social and nonsocial emotional and neutral stimuli. The Dot-Probe task was used to measure attention, and the Delayed-Non-Matching-to-Sample task with a 1s delay (DNMS) was used to measure recognition memory for these stimuli. Performance on the two tasks was examined during each of four month-long phases: Baseline, Lupron alone, Lupron+TE and Lupron+oil. It was predicted that T administration would lead to increased attention to negative social stimuli (i.e., negative facial expressions of unfamiliar conspecifics) and would improve memory for such stimuli. We found no evidence to support these predictions. In the Dot-Probe task, an attentional bias towards negative social stimuli was observed at baseline, but T treatment did not enhance this bias. Instead, monkeys had faster response times when treated with T compared to oil, independently of the emotional valence or social relevance of stimuli, perhaps reflecting an enhancing effect of T on reward sensitivity or general arousal. In the DNMS, animals had better memory for nonsocial compared to social stimuli and showed the poorest performance in the recognition of positive facial expressions. However, T did not affect performance on the task. Thus, even though monkeys were sensitive to the social relevance and emotional valence of the stimuli in the two tasks, T manipulations had no effect on attention or memory for these stimuli. Because habituation to the stimuli may have mitigated the effect of treatment in the attentional task, we suggest that T may increase attentional biases to negative social stimuli only during early exposure to the stimuli with acute treatment or when stimuli are highly arousing (i.e., dynamically presented) with chronic treatment. In addition, the data suggest that T does not enhance working memory for emotional stimuli in young male macaques.