Software measurement is a foundational research field in software engineering, and the COCOMOII is the most famous one in software measurement field. But after introducing COCOMOII into china from abroad, people have found that some places are not suitable for Chinese enterprises, and must make some improvements.This paper studies the characteristics of Chinese software industry, especially the characteristics of small-and-medium-sized enterprises which are in the majority of Chinese software industry. Their actual application cases in software measurement also take into account. The small-and-medium-sized enterprises' characteristics are frequent floating personnel, lack of high-quality talents, weakness of enterprise economics strength, and devoid of management standardization. This paper improves COCOMOII aiming at above characteristics, and brings forward a convenient and wieldy estimate flow. The improving model conquers former model's disadvantages, reduces difficulty of understanding and using, and decreases the workload because of cost estimating. Furthermore, the improving model enhances estimating veracity because it optimizes small-and-medium-sized enterprises' local project history data.Through validating the improving model, this paper proves that the improving model reaches its anticipative goal. The improving COCOMOII will provide a wieldy and veracious software measurement scheme, which will help to change the passive status of guessing only with experience to make software project management and layout. This model has great application worth at aspects of improving small-and-medium-sized enterprises' software project management, and advancing enterprise's benefit. |