Linearity is one of the most important specifications in electrical circuits.;In Chapter 1, a ladder-based transconductance networks has been adopted first time to build a low distortion analog filters for low frequency applications. This new technique eliminated the limitation of the application with the traditional passive resistors for low frequency applications. Based on the understanding of this relationship, a strategy for designing high linear analog continuous-time filters has been developed. According to our strategy, a prototype analog integrated filter has been designed and fabricated with AMI05 0.5 um standard CMOS process. Experimental results proved this technique has the ability to provide excellent linearity with very limited active area.;In Chapter 2, the relationships between the transconductance networks and major circuit specifications have been explored. The analysis reveals the trade off between the silicon area saved by the transconductance networks and the some other important specifications such as linearity, noise level and the process variations of the overall circuit. Experimental results of discrete component circuit matched very well with our analytical outcomes to predict the change of linearity and noise performance associated with different transconductance networks.;The Chapter 3 contains the analysis and mathematical proves of the optimum passive area allocations for several most popular analog active filters. Because the total area is now manageable by the technique introduced in the Chapter 1, the further reduce of the total area will be very important and useful for efficient utilizing the silicon area, especially with the today's fast growing area efficiency of the highly density digital circuits. This study presents the mathematical conclusion that the minimum passive area will be achieved with the equalized resistor and capacitor.;In the Chapter 4, a well recognized and highly honored current division circuit has been studied. Although it was claimed to be "inherently linear" and there are over 60 published works reported with high linearity based on this technique, our study discovered that this current division circuit can achieve, if proper circuit condition being managed, very limited linearity and all the experimental verified performance actually based on more general circuit principle. Besides its limitation, however, we invented a novel current division digital to analog converter (DAC) based on this technique. Benefiting from the simple circuit structure and moderate good linearity, a prototype 8-bit DAC was designed in TSMC018 0.2 um CMOS process and the post layout simulations exhibited the good linearity with very low power consumption and extreme small active area.;As the part of study of the output stage for the current division DAC discussed in the Chapter 4, a current mirror is expected to amplify the output current to drive the low resistive load. The strategy of achieving the optimum bandwidth of the cascode current mirror with fixed total current gain is discussed in the Chapter 5.;Improving the linearity of pipeline ADC has been the hottest and hardest topic in solid-state circuit community for decade. In the Chapter 6, a comprehensive study focus on the existing calibration algorithms for pipeline ADCs is presented. The benefits and limitations of different calibration algorithms have been discussed. Based on the understanding of those reported works, a new model-based calibration is delivered. The simulation results demonstrate that the model-based algorithms are vulnerable to the model accuracy and this weakness is very hard to be removed. From there, we predict the future developments of calibration algorithms that can break the linearity limitations for pipelined ADC. (Abstract shortened by UMI.)