Heroin addiction is a kind of chronic functional encephalopathy, which is the dependence on the psychoactive substance and accompanied with a series of physiological changes, such as impairment of cognitive function, weakening of adaptability and so on. Drug-related cues refer to the environmental stimuli that coexist with drug use, such as drug, action and tool, to name a few. Drug related cue-induced reactivity is a typical manifestation of addictive behavior and research has demonstrated that the attention bias of heroin abstainer to related cues is an important factor in predicting their re-inhalation. However, there are few studies about the heroin abstainer’s attention bias and specific differentiation of related clues. Do heroin abstainers have some cognitive processing and attention bias for direct cues (e.g., heroin drugs) and indirect cues (e.g., tools and actions)? That’s the goal of the present study. Previous studies on heroin addiction mainly used Stroop or dot-probe paradigm to explore the attention bias of heroin abusers, and ignored the conflicts of heroin abusers to different drug clues. In current study, we used dot probe paradigm combing with eye tracking technology to explore the attentional bias of heroin abstainers under different drug related cues. The present study used eye tracking to measure attention biases in specific component processes of visual attention (i.e., time to first fixation, first fixation duration, total fixation duration and total fixation count). 25 heroin dependents recruited from a domestic drug rehabilitation center in this study. We excluded a participates who could not complete the study. The mean age of participates were 41.92 (SD=4.98) and the duration of heroin using was 17.71 years (SD = 4.80). We conducted a 4 (picture classification: drugs, tools, actions, neutral) × 2 (probe position: consistent, inconsistent) within subject design to investigate the cognitive processing of heroin abstainers. The cues materials were divided into drug clues and neutral picture (D - N), tools clues and neutral picture (T-N), action clues and neutral (A - N) and neutral picture and neutral picture (N - N). The subjects were asked to press different keys on the keyboard as quickly and accurately as possible in response to the presented stimulus. At the same time, a SMI RED 250 Desktop eye-tracker (SensoMotoric Instruments, Germany) was used to record the eye movement data. The results showed that heroin abstainers had different attention bias patterns to different heroin related clues. Response time showed that heroin abstainers avoided paying attention to drug pictures, while maintained attention to tool and action pictures. Eye movement data showed that for heroin abstainers, the early detection to drugs was slowed down, but attention avoidance was maintained, and this attention model continued throughout the process on drugs. On the contrary, the early detections to tools and actions were accelerated, and attention to them was maintained continually. In summary, heroin abstainers’ attention bias to different drug-related cues was investigated. These findings provide behavioral and eye movement evidence for the hypothesis that heroin abstainers had different attentional bias patterns for different cue-related pictures. Because attention bias can reflect one's potential psychological state, understanding the attention bias mechanisms of heroin addicts provides important guidelines for revealing the addiction mechanism and detoxification treatment of heroin addicts.