The hypotesis has been formulated that the decison about generation of a creative imagery is made if a representation which is characterized by a low level of analogy with object from natural world is usefulness in the context of the task. However, the decision about generation of a reproductive imagery is made if a representation which is characterized by high level of similarity to the object form natural world is usefulness in the context of the task. Because there are different adaptation targets of reproductive and creative imagery, it was also assumed that surprising stimulus will affect the computation mechanism to produce the imagery with higher originality level and lower similarity level of imaginated objects to virtual objects than forseeable stimulus. On the other hand it can be assumed that forseeable stimulus are able to stimulate the computation mechanism to produce the imagery with lower originality level but higher similarity level than the surprising stimulus. Because the general intelligence as well as imagination are specific 'survival tools' and they increase subject's fitness, it is assumed to appear a postive correlation between the general (fluency) intelligence and adequacy of decisions made by the computation mechanism: in the conditions inducing generation of reproductive imagery it anticipated a positive correlation between the fluency intelligence and similarity of imaginated objects to ones existing in the virtual world but a negative correlation between the fluency intelligence and originality of the imagery; exactly the inverse relation was expected in the experimental conditions inducing generation of creative imagery. The first hypothesis was partially confirmed. The experiment didn't confirm relations of the fluency intelligence and the computation mechanism adequatness. The results of experiment were discussed from neuropsychological perspective and a necessity to use other kinds of measurement of the fluence intelligence than Raven's test in the future research was stressed.