To investigate a 9-year trend of fresh fruit and vegetable availability and factors associated with the net availability change in two contrasting neighbourhoods.

Longitudinal design. Data were collected in 2003, 2009 and 2012 through in-store observations. Fresh fruit and vegetable availability was presented by weight-adjusted counts of stores having designated varieties per 10 000 population.

A low-income minority neighbourhood and an adjacent middle-income racially mixed neighbourhood in Albany, NY, USA. These neighbourhoods became sites of fresh produce interventions after baseline data were collected.

A total of 111, 128 and 146 eligible food stores in respective years.

Fresh fruit availability (two or more varieties) increased in both neighbourhoods. Inventory expansion of existing stores and the convenience store intervention contributed to the significant increase (P for trend=0·04) of fresh fruit availability in the minority neighbourhood. Although not statistically significant (P>0·05), the availability of two or more dark-coloured fresh vegetables also increased in the mixed neighbourhood, but declined slightly in the minority neighbourhood. The secular (non-intervention) fresh vegetable availability rate ratio by neighbourhood reached 3·0 in 2012 (P<0·01). The net decline of fresh vegetable availability in the minority neighbourhood was primarily attributed to inventory reduction of existing stores.

Longitudinal observations revealed narrowed neighbourhood disparities of fresh fruit availability and widened gaps of fresh vegetable availability. Inventory shifts of existing stores impacted the net availability change more profoundly than store opening or closing in the minority neighbourhood. Findings support increasing the programme capacity of the convenience store intervention to address the fresh vegetable disparity.