The utilization of benzodiazepines (BZDs) and z-hypnotics has substantially increased with the aging of the population, but the risk of BZDs and z-hypnotics in the development of dementia remains a strong concern. This cohort study aimed to evaluate the risk of BZDs and z-hypnotics for subsequent dementia development with a special consideration of their half-lives and the concomitant use of these medications. People aged 65 years and older who were newly prescribed oral BZDs or z-hypnotics between 2003 and 2012 were identified from Taiwan’s National Health Insurance Research Database. All BZDs were categorized as long-acting drugs (≥ 20 h) or short-acting drugs (< 20 h) for further comparisons, and data were collected on a quarterly basis, starting on the first date of drug prescription and ending on the date of death, occurrence of dementia, or end of the follow-up period (December 31, 2012), whichever came first. All dementia events except vascular dementia occurring during the follow-up period were identified. Among 260,502 eligible subjects, short-acting BZDs and z-hypnotics users were at greater risk of dementia than long-acting users [adjusted odds ratio (95% confidence interval) in short-acting BZD users, 1.98 (1.89–2.07); z-hypnotic users, 1.79 (1.68–1.91); and long-acting BZD users, 1.47 (1.37–1.58)]. In addition, subjects concomitantly using 2 or more BZDs or z-hypnotics had a higher risk of dementia than those who used 1 of these drugs (4.79 (3.95–5.81)). The use of BZDs and z-hypnotics was strongly associated with the risk of dementia development, especially the short-acting BZDs, z-hypnotics, and concomitant use of multiple agents. These findings deserve further interventional studies for clarification.