Mandarin-speaking children with autism spectrum disorder (ASD) have difficulties producing aspect markers. The difficulties were explained in terms of pragmatic deficits since these children demonstrated strength in the comprehension of aspect markers using the Intermodal Preferential Looking (IPL) task. To verify whether this dissociation between production and comprehension could be replicated using another technique to the IPL, and if all children with ASD show difficulties in the production of aspect markers. A total of 34 children with ASD without cognitive delay, half with language impairment (ALI; mean age = 61.25 months old) and half with normal language (ALN, mean age = 61.52 months old), as well as 17 age-matched typically developing (TD) children (mean age = 61.38 months old) participated in a sentence-picture-matching task and a priming picture-description task to investigate their comprehension and production of Mandarin aspect markers zai-, -le and -zhe. In the comprehension task, children in the ALN group performed similarly to their TD peers, but those in the ALI group were less accurate on zai- and -le than TD children; children in all groups received higher accuracy when zai- was combined with Activity rather than Accomplishment verbs, and those in the ALI group was also more accurate when -le occurred with Achievement verbs, in contrast to Activity verbs. In the production task, children in the ALI group produced fewer targets and more irrelevant sentences with zai- than their TD peers, and they tended to produce bare verbs for -le and -zhe than TD children; children in all groups tended to combine zai- with Activity verbs, and those in the ALN group also tended to combine -le with Achievement verbs. The comprehension and production of Mandarin aspect markers by children with ASD are linked to general language abilities, and interactions between lexical and grammatical aspect. Patterns of performance are similar to those of TD peers only for the subgroup with spared global language, while pragmatic deficits are pervasive throughout the spectrum. Therefore, training on formal language, with a specific emphasis on aspectual rather than pragmatic abilities, may be more effective at enhancing the production of aspect markers. What is already known on this subject Mandarin-speaking children with ASD have difficulties producing aspect markers but demonstrate strengths in aspectual comprehension via the IPL task. Therefore, it has been proposed that their 'specific' difficulties in aspectual production should be ascribed to their pragmatic deficits. However, pragmatic deficits are highly pervasive in children with ASD while only a subgroup of children with ASD who are impaired in language development (children with ALI) show difficulties in producing tense/aspect morphology. Pursuing this reasoning, pragmatic deficits might not be the critical factor impacting the performance of children with ASD in aspectual production. What this study adds Children with ASD were divided into one group with ALI and the other with normal language (ALN). Results of a sentence-picture-matching and a priming picture-description task illustrated that both groups preserved the comprehension of Mandarin aspect markers zai-, -le and -zhe. However, children with ALI performed worse than age-matched TD children, while children with ALN demonstrated similar performance to TD children in aspectual production. These findings, coupled with the fact that pragmatic challenges affect individuals throughout the spectrum, suggest that general language abilities rather than pragmatics better explain the performance of children with ASD on aspectual production. What are the potential or actual clinical implications of this work? Since general language abilities rather than pragmatic deficits of children with ASD determine their performance on the production of aspect markers, direct training on the use of aspect markers or more global language therapy could benefit children with ASD in the production of aspect markers.