When analyzing longitudinal data, a typical assumption is that all individuals follow a single trajectory over time. In the presence of substantial between-subject heterogeneity, it may be more realistic to assume that individuals follow more than one trajectory. Latent class methods for longitudinal data are a collection of methods that assume individuals in a sample belong to one of several latent trajectory classes rather than a single class. There are several advantages to this approach including an improved ability to capture between-subject variability and a better understanding of how individual sub-groups change over time. In this talk we provide an introduction to latent class methods for longitudinal data and review several different methods for estimating these models. To fix ideas, we describe an analysis of blood pressure trajectories from 4681 participants in the CARDIA study collected over 25 years.