Nowadays CMOS active pixel sensor (APS) cameras are widely used in video applications. APS consists of photodiode and some control transistors. During integration time photogenerated signal charges are collected by the photodiode. The main problem for CMOS active pixel sensors (APS) design is its fill factor and photosensitivity improvement. The fill factor is the aim function of optimization in CMOS pixel design. The constraints, which are imposed on the function, are: reset and read-out times delays, signal-to noise ratio, values of horizontal and vertical modulation transfer function (MTF), pixel sizes, and project rules. Analytical models are used in order to describe APS features. 2D numerical models for process, potential, and charge distributions simulation have been used for calibration of analytical model coefficients. Small changing of the process parameters has allowed getting required charge capacity and antiblooming possibility for the photodiode. Integrated signal charge has reduced effective depth of photodiode potential well. Using pixel layout the geometrical model has been developed. The model takes into consideration sizes and layouts of the pixel components: photodiode, transistors, buses, and also project rule. The possible CMOS APS design method that allows maximization of fill factor and photosensitivity has been presented. Using developed CMOS APS models we have solved this problem as mathematical optimization task. As example, we have considered APS with 4 control transistors.