Computer modelling of personality and behaviour is becoming increasingly important in many fields of computer science and psychology. Personality and emotion-driven Believable Agents are needed in areas like human-machine interfaces, electronic advertising and, most notably, electronic entertainment. Computer models of personality can help explain personality by illustrating its underlying structure and dynamics. This work presents a neural network model of personality and personality change. The goals are to help understand personality and create more realistic and believable characters for interactive video games. The model is based largely on trait theories of personality. Behaviour in the model results from the interaction of three components: (1) personality-based predispositions for behaviour, (2) moods/emotions and (3) environmental situations. Personality develops gradually over time depending on the situations encountered. Modelling personality change produces interesting and believable virtual characters whose behaviours change in psychologically plausible ways.