Henrietta Leavitt was born the daughter of a Congregational minister in Lancaster, Massachusetts. Her interest in astronomy was aroused while she was at Radcliffe College (then the Society for the Collegiate Instruction of Women), from which she graduated in 1892. In 1895 she became a volunteer research assistant at the Harvard College Observatory, receiving a permanent post in 1902. She was soon head of the department of photographic photometry. Like her colleague Annie Cannon, she was extremely deaf.

Leavitt's work involved the determination of the photographic magnitudes of stars, i.e., their brightness as recorded on a photographic plate. The photographic magnitude of a star differs somewhat from its visual magnitude since a photographic emulsion is more sensitive to blue light than the eye. The accurate measurement of visual magnitudes had been part of the program of the Harvard College Observatory since the 1870s. In 1907 the director of the observatory, Edward Pickering, announced plans to redetermine stellar magnitudes by photographic techniques. The photographic magnitudes of a group of stars near the north celestial pole were to act as standards of reference for other stars. Leavitt was selected to measure these magnitudes, known as the ‘north polar sequence’, and the results were published in 1917 in the Annals of Harvard College Observatory (vol. 71 no. 3). She also spent many years measuring secondary stellar magnitudes, based on the north polar sequence, which was adopted as an international standard until superseded by photometric measurements of magnitude.

Leavitt also did much work on variable stars, discovering about 2400 – roughly one half of those known in her time. She is best known, however, for her studies of Cepheid variables. At Harvard Observatory's field station at Arequipo, Peru, a series of photographic plates had been taken of the Magellanic Clouds (now known to be small neighboring galaxies). From her analysis of the plates, Leavitt detected nearly 1800 variable stars, some of which belonged to a class known as Cepheid variables. The variation in brightness of Cepheids is extremely regular and in 1908 Leavitt noted that the brighter Cepheids had the longer periods. By 1912 she was able to show that the apparent magnitude, i.e., observed brightness, of Cepheids decreased linearly with the logarithm of the period.

It was this seemingly simple discovery that led to an invaluable means for determining very great distances; previous to this only distances out to a hundred light-years could be estimated. Leavitt's work on the light variation of Cepheids was extended first by Ejnar Hertzsprung and Harlow Shapley and then by Walter Baade to give the period–luminosity relation of Cepheids. Using this relation the luminosity, or intrinsic brightness, of a Cepheid can be determined directly from a measure of its period and this in turn allows the distance – of the Cepheid and its surroundings – to be calculated. Distances of galaxies up to ten million light-years away can be determined this way.