The present study aimed to develop a Chinese equivalence version of the Letter-Number (LN) Span Test and to explore the preliminary construct and discriminative validity of the developed version among a group of healthy Chinese volunteers and patients with stroke. A group of 165 (73 men and 92 women) healthy participants were recruited for the validation study, most of them were undergraduates or postgraduates. Moreover, a comparison was made between nine patients with stroke and the healthy controls. For the healthy sample, the Chinese version correlated significantly with the English version in total number of correct span (r = .6, p < .00001) and the longest span (r = .5, p < .00005). The Chinese version of LN Span Test was also found to be significantly associated with memory-loaded tests but not other tests. A series of ANCOVAs controlling for age, education, and IQ indicated that stroke patients performed significantly worse than the healthy controls in LN Span total number of correct responses (p < .04), immediate recall (p < .0005), and delayed recall (p < .0005) of WMS-R, SART total number of correct response (p < .0005), PASAT dyads correct response (p < .01). The preliminary findings suggest that the Chinese version of the LN Span Test shows impressive preliminary validity among a group of healthy volunteers and an impressive clinical discriminative validity among a group of stroke cases.