Colour is often claimed to be the most emotional appearance factor of a product. Therefore, many researchers have been conducted researches on this factor to obtain the relationship between colour and emotion. In the real life, colour are always seen in combination with other colours and it is virtually impossible to see colours individually. Even in a psychophysical experiment and under a light cabinet the samples are observed in combination with the background colour, which is normally a gray with L*= 50. On the basis of appearance models, the colour of a sample is affected by the colour of neighbourhood samples, and consequently, the colour emotion imposed by that sample can be changed, accordingly. On the other hand, the Cupchik model considers two levels of emotional responses: reactive and reflective. The reactive level describes the emotional responses to the configurations of features in visual arts and in this level, apart from the product application, a relationship between emotions and colours are available. However, in the reflective level, the application of the product affects the human emotional response. The aim of the present research is to clarify the relationship between colour pairs and the emotion they may produce and to investigate the effect of culture, gender, age and education on colour emotion. To do this, a psychophysical experiment was carried out to investigate colour emotion for two-colour combination. Five colour emotion scales, heavy/light, warm/cool, active/passive, like/dislike and harmony/disharmony, were used to assess 190 colour pairs presented on a calibrated LCD in a darkened room. 51 Iranian observers with normal vision, examined with Ishihara colour blindness test, attended the experiment. Experimental results showed that the two scales like/dislike and harmony/disharmony have close correlation. The results of the experiments showed that "like-dislike" and "harmonious-disharmonious" colour emotion scales have close correlation. Also, according to the results obtained from the correlation coefficients, it seems that there is no gender, culture, environment, educational background and age effect on the three colour emotion scales, which make the Li Chen Ou's colour emotion space, i.e. "heavy-light", "warm-cool" and "active-passive". However, apart from the gender, these factors affect the response of the observers to the "like-dislike" scale and it means colour preference is likely to be affected by these factors. The insignificant effect of the gender on colour preference is not in accordance with the results obtained from previous studies and this disagreement demands further work. Also, the results indicated that Iranian prefer the colour pairs which have small hue difference and those for which the average of their chroma is high. The preference of the observers who have design education background for the colour pairs with lower mean lightness value or similar hues is more than the other observers. Moreover, it was observed that male observers prefer colour pairs with higher mean chroma more than what the female observers prefer. The results, also, indicated that the 20-29 years old observers prefer colour pairs with higher mean chroma or higher hue difference more than what the observers in 30-47 years old age group prefer. The application of the Li Chen Ou's proposed models of colour emotion for the Iranian results showed that there is no good prediction for the "heavy-light", "like-dislike" and "harmonious-disharmonious" scales, however, the models have fairly good prediction for the "warm-cool" and "active-passive" scales. Key Words: Colour Emotion, Colour Appearance, Colour Combination, Colour Harmony, Colour preference.