Importance Long-term exposure to total fine particulate matter (PM 2.5 ) is a recognized dementia risk factor, but less is known about wildfire-generated PM 2.5 , an increasingly common PM 2.5 source. Objective To assess the association between long-term wildfire and nonwildfire PM 2.5 exposure and risk of incident dementia. Design, Setting, and Participants This open cohort study was conducted using January 2008 to December 2019 electronic health record (EHR) data among members of Kaiser Permanente Southern California (KPSC), which serves 4.7 million people across 10 California counties. KPSC members aged 60 years or older were eligible for inclusion. Members were excluded if they did not meet eligibility criteria, if they had a dementia diagnosis before cohort entry, or if EHR data lacked address information. Data analysis was conducted from May 2023 to May 2024. Exposures Three-year rolling mean wildfire and nonwildfire PM 2.5 in member census tracts from January 2006 to December 2019, updated quarterly and estimated via monitoring and remote-sensing data and statistical techniques. Main Outcome and Measures The primary outcome was incident dementia, identified using diagnostic codes in the EHR. Odds of dementia diagnoses associated with 3-year mean wildfire and nonwildfire PM 2.5 exposure were estimated using a discrete-time approach with pooled logistic regression. Models adjusted for age, sex, race and ethnicity (considered as a social construct rather than as a biological determinant), marital status, smoking status, calendar year, and census tract–level poverty and population density. Stratified models assessed effect measure modification by age, sex, race and ethnicity, and census tract–level poverty. Results Among 1.64 million KPSC members aged 60 years or older during the study period, 1 223 107 members were eligible for inclusion in this study. The study population consisted of 644 766 female members (53.0%). In total, 319 521 members identified as Hispanic (26.0%), 601 334 members identified as non-Hispanic White (49.0%), and 80 993 members received a dementia diagnosis during follow-up (6.6%). In adjusted models, a 1-μg/m 3 increase in the 3-year mean of wildfire PM 2.5 exposure was associated with an 18% increase in the odds of dementia diagnosis (odds ratio [OR], 1.18; 95% CI, 1.03-1.34). In comparison, a 1-μg/m 3 increase in nonwildfire PM 2.5 exposure was associated with a 1% increase (OR, 1.01; 95% CI, 1.01-1.02). For wildfire PM 2.5 exposure, associations were stronger among members less than 75 years old upon cohort entry, members from racially minoritized subgroups, and those living in high-poverty vs low-poverty census tracts. Conclusions and Relevance In this cohort study, after adjusting for measured confounders, long-term exposure to wildfire and nonwildfire PM 2.5 over a 3-year period was associated with dementia diagnoses. As the climate changes, interventions focused on reducing wildfire PM 2.5 exposure may reduce dementia diagnoses and related inequities.