Abstract： Sentences such as “Someone shot the servant of the actress who is on the balcony” were ambiguous, because the internal clause “who is on the balcony” was either attached to the non-local noun phrase “the servant”, which is called high attachment preference, or attached to the local noun phrase “the actress”, which is called low attachment preference. Currently available evidences in the literature demonstrated considerable cross-linguistic differences in relative clause attachment preferences in on-line and off-line processing. Different models of ambiguity resolution have been suggested to explain the cross-linguistic relative clause attachment preference. One of the most widely accepted accounts of attachment preference were the Parameter Variation Model developed by Gibson (1996), which held that ambiguity resolution in sentence processing was guided by two principles: Recency and Predicate proximity. Investigating the mechanism of ambiguity resolution in sentence processing is more helpful for exploring the inner processing mechanism of human language. 170 participants were sampled for our experiment. Using off-line subjective ratings and eye-movement tracking, this paper investigated the influence of cognitive style on the preference of syntactic ingredients attachment by comparing two groups of participants’ performances of ambiguity resolution in sentences processing. One group was identified with the field dependence in cognitive style, while the second group was the field independence. They were all native speakers with normal or corrected to normal insight, right handed. Materials used in the study were classified into three categories: sentence with ambiguity in attachment, sentence with high attachment preference, and sentence with low attachment preference. The results from off-line study showed that, in Chinese, readers preferred to attach the segments to the final noun phrase in sentence. In particular, this preference was not varied with the variation of the experimental materials. Results from the eye-movement tracking study showed that, cognitive style has significant impacts on the methods used to resolve the ambiguity in attachment. Specifically, readers with dependent cognitive style have more difficulties in resolving the ambiguity at regions of NP2, which was reflected in eye-movement indexes, such as the first fixation, the first run duration, the regression duration, and the total time. Additionally, we found significant differences in the three types of sentence processing. In general, sentences with ambiguity in attachment were most difficult to comprehend, and sentences with low attachment preference were easiest to process. In sum, participants with independent cognitive styles have more difficulties in resolving the ambiguities in syntactic ingredients attachment. Two opinions regarding the difficulties in ambiguity resolution for participants with independent cognitive styles were suggested. One argued that individuals with independent cognitive styles were inferior in inhibiting interference from other syntactic fragments in sentence processing. The other taught that the limited capacity in working memory was the key factor to impede individuals with independent cognitive styles to address syntactic ambiguity resolution with success.